BEHAVIORAL AND BRAIN SCIENCES (1993) 16, 115-186

Printed in the United States ot America

A perspective for viewing the

history of psychophysics
David J. Murray
Department of Psychology, Queen's University,
Kingston, Ontario, Canada K7L 3N6
Electronic mail: murrayd@qucdn.queensu.ca

Abstract: Fechner's conception of psychophysics included both "outer psychophysics," the relation between stimulus intensity and
the response reflecting sensation strength, and "inner psychophysics,' the relation between neurelectric responses and sensation
strength. In his own time outer psychophysics focussed on the form of the psychophysical law, with Fechner espousing a logarithmic
law, Delboeuf a variant of the logarithmic law incorporating a resting level of neural activity, and Plateau a power law. One of the
issues on which the dispute was focussed concerned the appearance of contrasts if the overall illumination was increased or decreased;
another issue was the question of whether a sensation of a "just noticeable difference" established for one value of a sensory dimension
appeared the same for a value elsewhere on the dimension. The development of "inner psychophysics" led through the works of
Delboeuf, Solomons, Jastrow, and Thurstone to modern signal detection theory. A third line of research, devoted to the question of
what was meant by the "measurement" of sensation strength, stemmed from the criticism of Fechner's work by von Kries (1882) and
others. Although a valid body of science could be built up without the intervening variable called "sensation strength," such a science
might be a cumbersome representation of reality. When an optical contrast is set up, and its overall illumination is increased or
decreased, subjective contrasts involving medium levels of lightness vary little as illumination varies (as a power law based on
sensation ratios or a logarithmic law based on sensation differences predict), but subjective contrasts involving extreme levels of
lightness might be subject to the effects of other variables.

Keywords: Delboeuf; distinctiveness; Fechner; history of psychology; measurement; philosophy of science; psychophysics;
sensation; Stevens

1. History deavour and in particular the status of so-called internal

The purpose of this target article is to provide a perspec-
tive for viewing the history of psychophysics. There are a
number of reasons for its appearance at this time. One is
the revival of interest in the early history of psycho-
physics, partly due to the centenary in 1987 of Fechner's
death; scholars like Scheerer (1987) have shown that
many of Fechner's original ideas about psychophysics
have been underreported and therefore undervalued.
Another is that psychophysics is the topic of two impor-
tant recent reviews, those of Laming (1986 - see also
multiple book review, BBS 11(3) 1988) and Falmagne
(1985). The area seems to be dividing into two camps, one
leaning towards an integration of psychophysical data and
theory with the findings of sensory physiology, as repre-
sented by Laming, the other attempting to integrate
psychophysics into "measurement theory" as charac-
terised by the work of Falmagne. A third reason is
Kruegers (1989a) recent target article in Behavioral and
Brain Sciences, which stirred up a stimulating degree of
controversy, indicating a lack of unity among psycho-
physicists on some of the most fundamental questions
such as the nature of the psychophysical law and the
feasibility of mental measurement. A fourth reason is that
psychophysics raises many issues of interest to all experi-
mental psychologists concerning the nature of their en-
scalable magnitudes such as sensation strength (Murray
1987).
The gist of this paper is schematized in Figure 1: Is it a
valid representation of what has happened over the last
100 years or so? Certain issues arose as I developed
Figure 1 that I believe merit more public discussion;
these will be presented in sections 2.1 and 2.2 of this
article. My main claim is that issues raised originally in
the nineteenth century, in work that is not very widely
known, remain central to psychophysics today and that
pertinent and fruitful progress has been made. I hope to
clarify the extent to which the task needs finishing. I shall
accordingly offer a historical narrative stressing the con-
nections between the writings of Fechner and those of
certain modern psychophysicists, followed by some gen-
eral remarks on two important issues in the early history
of psychophysics: (1) the role of sensation strength in
psychological science and (2) the appearance of contrasts
under changing illuminations. My conclusion will be that
Fechner himself was more aware of the true nature of
psychophysical enquiry than we usually realize: He had a
philosophical sophistication on scientific matters that has
been underestimated in the past, partly perhaps because
of the mystical origins of some of his theorizing. On the
other hand, some of his well-known and most basic ideas
have had to be reformulated.

1993 Cambridge University Press 0140-525X193 $5.00+.00 115

Murray: History of psychophysics
Stevens 1957.1961
Falmigne 1983
I time is measured by its energy P (energy understood in
Knicgn 1989 LUM 4 NM
Figure 1. A chart showing how the history of psychophysics
can be organised in relation to Fechner's distinction between
"outer psychophysics" and "inner psychophysics." At the top the
chain / * E * S >flrefers to the chain physical intensity *
neurelectric response * sensation * overt response. Some
important reviews of psychophysics, including those of G. E.
Muller(1878), Kohler (1886), Titchener (1905), and Gescheider
(1976), are not shown. The chart should not be interpreted as
implying that the individuals mentioned are more important
than others not shown; some modern investigators have contro-
versial ideas and some major modern contributors to psycho-
physics, such as R. D. Luce and D. M. Green, have contributed
to all the columns.
1.1. Fechner's conception of the psychophysical law
Scheerer (1987) has pointed out that Fechner, like
Wundt, was exceptionally prolific and that only one of his
writings, the first volume of the Elements of psycho-
physics (1860/1964) was available in translation (Fechner
1860/1966) until recently. Scheerer himself has trans-
lated three further passages: (a) an appendix from the
early mystical work, the Zend-Avesta (1851) in which
Fechner makes it quite clear that he intuited the log-
arithmic law relating sensation magnitude to stimulus
magnitude long before he did any experiments or even
knew about Weber's Law (see Fechner 1987a); (b) a
passage from the Revision der Hauptpuncte der Psycho-
physik (1882), a late update on the Elements, in which
Fechner showed a sophistication concerning the nervous
system that allowed him to offer a model of memory which
strongly anticipates modern connectionist views such as
those of Rumelhart, McClelland, and the PDP Research
Group (1986) - we shall not discuss this here; and (c) a
passage from his last article (Fechner 1887; see 1987b) in
which Fechner makes it clear that he believed that
sensation magnitude can be scaled and in particular that
"difference sensations" (a sensation arising from a cogni-
tion that two stimuli differ from each other) can be
measured and used to derive a scale of "sensation differ-
ences" (numbers reflecting differences between the
strengths of two sensations). "Difference sensations" are
better estimated, he argued, from "summated jnd" mea-
surements or "bisection" measurements than from magni-
116 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
tude estimations. This last statement is probably impor-
tant when viewed in the light of an article by von Kries
(1882), which attacked Fechner's enterprise at its very
heart: Von Kries denied that sensation magnitude could
be "measured" according to the criteria of "measurement"
laid down by physics. Because Fechner felt obliged to
defend himself against this and other attacks, he contin-
ued to write about psychophysics for another 27 years
after the appearance of the Elements. I hope to show that
some of his views, such as those on "inner psychophysics,"
were more defensible than others, such as the logarithmic
psychophysical law.
Fechner's Elements of psychophysics elaborated on the
argument first put forward in the Zend-Avesta, in which
Fechner wrote as follows:
If the strength of the physical activity actually under-
lying some mental activity at some point in space and
Notwichl987
the sense of mechanics), and if its change, assuming an
infinitely small part of time and space, is named d$,
then the accompanying change in the intensity of the
mental activity, to be estimated by feeling or in con-
sciousness, is not proportional to the energy change
dp, but to the relative change dp/p. . . . (Fechner
1987a, p. 203)1
The term dp is clearly a differential. Fechner was still
to discover Weber's Law according to which an increment
in p, in order to be just noticeably different from P, had to
be a constant fraction of p. However, when Fechner did
read Weber's books of 1834 and 1846 (see Weber 1978),
he speculated that what he had said about the mental
activity as dependent on the ratio of dp to P might be
related to the discovery that a just noticeable difference in
sensation depended on the ratio of the increment in P to P
itself. By 1860 he had clearly made the connection, as is
evident from the following equation (here he uses y to
signify sensation strength):
According to the experimentally demonstrated law
of Weber, dy stays constant if dp/p remains constant no
matter what absolute values dp and P take; and accord-
ing to a mathematical auxiliary principle that seems to
be valid a priori, the changes dy and dp remain propor-
tional to each other so long as they remain very small.
(Fechner 1860/1964, vol. 2, p. 10)
Because the modern abbreviation for physical intensity
is usually 7 rather than P, which has a special meaning in
signal detection theory, we shall here change the notation
so that "7" refers to physical intensity. If A7 is the incre-
ment in 7 needed to elicit a judgment that 7 + A7 is just
noticeably different from 7, then Weber's Law says that
A/ = kl (1)
where k is a positive constant. In the Elements Fechner
amended the argument of the Zend-Avesta so that now an
increment in 7 that led to a just noticeable difference
between 7 and 7 + A7 was taken as the starting point for
the derivation of a psychophysical law. If A7 is this
increment and S is sensation strength then a just notice-
able increase in sensation strength AS will be a function of
the relationship of A7 to 7: This "fundamental formula' is
AS = C(A ///) (2)
where C is a positive constant. Note that Krueger (1989a;
1989r) prefers to replace C with j/k, where j is the

(constant) subjective magnitude of a just noticeable differ-
ence. In the Elements, Fechner went on to apply calculus
to Equation 1, as follows. Dividing both sides of Equation
2 by A/ gives
AS/A/ = C(l/1) (3)
Fechner (Elements, vol. 2, p. 9) then makes the as-
sumption that, just as a physically small quantity A/ can
be reduced without limit to the differential dl, so a small
"quantity" of sensation can be expressed, in the limit, as
dS. It may be noted that this assumption has been
challenged by Luce and Edwards (1958) on formal mathe-
matical grounds. 2 Fechner then proposes the differential
equation (taking limiting values of A/ and AS) that follows
from Equation 3:
dS/dl = C(l/I)
The standard solution is S = C In / + constant of
integration. Fechner showed that this in turn reduced to
S = C In (///) (4)
where / 0 is the "absolute threshold" value of/,- S = 0 (no
sensation) where / = l0. In order to apply this logarithmic
psychophysical law to real life, Fechner suggested that we
could estimate both Zo, and A/ in the laboratory by various
psychophysical methods. These include the method of
adjustment and the method of "right" and "wrong" cases:
We shall return later to a discussion of how these fit into a
more general scheme concerning methods. For Fechner,
the analytical arguments could only take on life if the
psychophysical methods could be applied in the labora-
tory; the study of the advantages and drawbacks of the
various methods came to dominate psychophysics partic-
ularly in the period between 1890 and 1920 - Titchener's
monumental Experimental psychology (1901-1905) is an
exhaustive account of the early literature on Fechner's
methods.
In developing his theory, Fechner considered that
since Weber's Law formed a cornerstone from which he
could move forward to derive his logarithmic psycho-
physical law by a logical procedure, it was of the greatest
importance to establish the validity of Weber's Law. To
do so required accurate measurements of the differential
threshold, and this was the incentive for Fechner to
develop his psychophysical methods; furthermore, one of
his ways of demonstrating that Weber's Law might be
valid was to take various values of I and show that the
percentage variability ofjudgments concerning whether I
+ A/ was different from 7 remained roughly constant for
all values of I (Fechner 1860/1966, pp. 85-93). This
treatment of variability involved an analysis of properties
of the normal distribution. (Later, in his Kollektiv-
masslehre (1897), Fechner developed a number of other
statistical ideas, including the evaluation of data whose
distribution was skewed and of whether in a series of
measurements, each measurement was independent of
the previous one; see also Heidelberger 1987). Fechner's
own experimentation, as reported in the Elements, was
largely restricted to the determination of thresholds using
weights and line lengths; perhaps because of the expense
of engraving figures, nowhere in any of his three books on
psychophysics, the Elements (1860), In Sachen der Psy-
chophysik (1877), or the Revision der Hauptpuncte der
Psychophysik (1882), does Fechner give a graph or a table
Murray: History of psychophysics
showing sensation magnitudes as a logarithmic function of
stimulus intensity. It can also be claimed that the calcula-
tion of sensation magnitudes was unnecessary given the
presentation of a formula for calculating them. Falmagne
(1986, pp. 1-10) has suggested that the actual construc-
tion of a scale of sensation magnitudes based on estimates
of jnds would necessarily be approximate because errors
made in the determination of an initial jnd would have a
cumulative effect in a scale constructed of concatenated
jnds.
Fechner also devoted a great deal of space to discussing
criticisms of Weber's Law: These criticisms, and Fech-
ner's rebuttals, were reviewed in the nineteenth century
by Kohler (1886) and more recently by Murray and Ross
(1988), who also assess Fechner's use of the law in a
variety of contexts. The criticisms resulted from work by
Aubert on brightness discrimination, by Hering on
weight discrimination, and other work which in general
led Fechner to conclude by the end of his life that
Weber's Law was valid, except for extreme values, for
prothetic (intensitive) dimensions like brightness and
loudness, as well as length and time discrimination; it
might not be valid for metathetic dimensions like hue and
pitch, in which sensory quality as well as "quantity"
changes.
1.2. Fechner's "Inner psychophysics"
Fechner made an important claim about the logarithmic
function which was apparently shrugged off by Hering
and other investigators of his time on the grounds that it
was too metaphysical (Scheerer 1992); he proposed that
the transformation of physical intensity into sensation
magnitude, which had to be compressive in view of the
logarithmic nature of the transformation, took place at the
time physical events in the brain were translated into
mental events of sensation. In order to understand this
view, we have to understand Fechner's distinction be-
tween "outer" and "inner" psychophysics. (Here, "inner"
is a translation of the German word "inner" but "encapsu-
lated" might be an alternative translation.) Outer psycho-
physics studied the relationship between the observable
physical stimulus and the reported psychological re-
sponse; for example, the psychophysical law relates / to S,
both of which are based on observable data. It was
Fechner's view that the most reliable estimate of S was
obtained from a summated jnd scale or a bisection scale
and he had little respect for what we would now call
"direct magnitude" estimates of S. He felt that these had
no more value than ratings of "beauty" or other subjective
impressions. Nowadays we adopt the distinction between
"confusion scales" (such as summated jnd scales or paired-
comparison data), "partition scales" (such as bisection data
or category judgments), and ratio scales (such as are
obtained from direct magnitude estimates or production).
Fechner might have found this nomenclature inappropri-
ate: For him, more "confusion" would have been associ-
ated with magnitude estimates. Moreover, if one insisted
on making a dichotomy between "direct" and "indirect"
measures (where the reality may be that we are dealing
with a continuum) some individuals may consider parti-
tion scales to be direct rather than indirect. But, since the
data of outer psychophysics in all these cases were based
on measurement, even though nominally indirect as in
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 117
Murray: History of psychophysics
the case of a summated jnd scale, outer psychophysics was
the type of psychophysics most easily carried out.
Inner psychophysics, however, was just as important
for Fechner. Inner psychophysics concerned the relation-
ship between what we would now call neurelectric as-
pects of the stimulus and the mental experience thereby
evoked: At the heart of inner psychophysics was the
"mind/brain" problem - Scheerer (1992) has remarked of
Fechner's opinion: ". . . to use terms not employed by
Fechner himself, the mind is an emergent property of the
entire psychophysical system and cannot be reduced to
one of its (functional or spatial) components" (p. 8).
Essentially Fechner thought of a physical stimulus as
influencing the nervous system by imposing a set of
ripple-like oscillations on a resting state of high-
amplitude low-frequency oscillations: The oscillations im-
posed by the physical stimulus had to be discriminable
from the resting state and furthermore had to be of such a
degree of discriminability as to cross a particular "thresh-
old" if the subject was to become conscious of the stim-
ulus. It was quite possible, in his view, for activity from a
sensory receptor to influence brain activity without influ-
encing the activity of consciousness, and as early as the
Zend-Avesta of 1851, Fechner speculated that the
amount of "work" necessary for a stimulation to become
conscious could be reflected in a number. A conscious
sensation could have a number ascribed to its magnitude
(a number predicted by Fechner's Law): If nervous activ-
ity was aroused by a stimulus / that was below the 70
needed to produce a sensation, the nonexistent sensation
that could not be aroused was nevertheless given a nega-
tive number. These negative numbers had no meaning in
reality except as an index of how subthreshold a level of
activity was; in later years Fechner fretted at the way
people assumed he was talking about something called
"negative sensations" as if they had an ontological status
equivalent to real sensations, when he never meant to
imply this at all (this issue is discussed at greater length by
Murray 1990).
Thus, the activity induced by a physical stimulus in the
nervous system had to cross several "thresholds" of ner-
vous activity before a particular threshold was reached at
which it was translated into conscious experience; it was
at that point that Fechner localized the transition from
many to one, from a broad range of physical intensities to
a narrow range of discriminably different sensations. This
transformation was nonlinear - it was at this junction,
rather than at the periphery, that Fechner localized the
logarithmic transformation. Three other aspects of Fech-
ner's writings on inner psychophysics may be mentioned
in passing: (1) Once a sensation had become conscious, it
was possible by "selective attention" to concentrate on
something else and thereby be temporarily unconscious
of that sensation; (2) as mentioned earlier, the nervous
activity referred to was held to occupy the whole brain,
with memory activity being aroused by a sort of reso-
nance; (3) since the nervous activity evoked by a stimulus
was added to the preexisting level of activity, this implied
that for Fechner all stimulus-related nervous activity had
to be measured in relation to other nervous activity. It
would then follow, because of the property that conscious
experience depended upon nervous activity, that individ-
ual conscious experiences also had to be discriminated
from a background level. Scheerer (1987) briefly indicates
118 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
that this view anticipated signal detection theory; how-
ever, signal detection theory had a separate history,
starting with Neyman and Pearson's (1928) analysis of the
process of statistical decision making. Moreover, signal
detection theory is more explicit than was Fechner on the
need to distinguish "internal noise" in the observer from
"external noise" present in the stimulus (for example, the
quantum or particle nature of light insures that the
stimulus has intrinsic variability). It can also be claimed
that the interpretation of the word "noise" as neurological
noise is only one way to handle it within the framework of
the signal detection model of decision making. However,
since signal detection theory is an intrinsic part of the
history of psychophysics it is included in Figure 1 in the
context of its logical relationship to Fechner's inner psy-
chophysics, even though a historical connection is diffi-
cult to trace.
It may also be noted here that, in assuming that a
nervous excitation had to exceed a particular nervous
threshold before a conscious perception concerning a
difference in sensation was possible, Fechner came close
to asserting something like Stevens's later neural quan-
tum theory: Corso (1967) explicitly stated that Fechner
and others "argued for the noncontinuity notion which
included the threshold concept" (p. 423). This raises the
difficult question of how Fechner would have viewed
subsequent work according to which the mere existence
of variability in data entailed that there was a variability in
the nervous system that would make the concept of a fixed
threshold nonviable. Later, we shall review evidence on
the line of research stemming from inner psychophysics
according to which Thurstone and the signal detection
theorists made variability a datum in the derivation of
ideas about difference sensations; and, as mentioned, we
know that Fechner himself was interested in the mea-
surement of variability. So this is perhaps the place to say
that, although Fechner foreshadowed signal detection
theory in the important respect that he considered signals
to have "distinctiveness" only in the context of a back-
ground of noise, he may well have disapproved of any
reliance on the variability of data in order to measure that
"distinctiveness." Stout (1989) has shown how even in the
1920s psychologists did not always see eye-to-eye in the
meaning to be ascribed to the variability of data.
To summarize Fechner's work so far, we may say that
he intuited that sensation magnitude was a logarithmic
function of physical intensity; an appropriate demonstra-
tion would show that, once Weber's Law was validated, a
scale of sensation magnitude could be based on a knowl-
edge ofjust noticeable differences. Fechner (1987b) him-
self realized, however, how the construction of a scale of
concatenated jnds could be carried out independent of
Weber's Law. He noted that "measurement of difference
sensations is by no means restricted to the validity of
Weber's Law. Indeed . . . in principle it has nothing to
do with it" (p. 215). Moreover, Luce and Edwards (1958)
have argued that since in the construction of a scale of
magnitude jnds are concatenated, this could be carried
out even if Weber's Law were invalid, a point also made
by Krueger (1989a). And Parker and Schneider (1980)
claimed that although Weber's Law was indeed invalid for
judgments of loudness, a scale of subjective loudness
could nevertheless be constructed on the basis of jnds.
Fechner also suggested that the locus of the logarithmic

transformation was the junction of the brain/mind inter-
face. He disliked the idea of measuring sensory magni-
tude by ratings. He felt that a full discussion of sensory
magnitude would involve a consideration of preexisting
levels of neural activity, and he maintained that sensory
magnitude was a measurable entity. Our main contention
will be that the history of psychophysics since Fechner's
time can be seen as having two broad strands, one
resulting from his activity on outer psychophysics - a
strand that included unexpectedly strong criticism both of
his belief that the psychological law was logarithmic and of
his belief in the measurability of sensory magnitude and
a second strand, logically connected with his inner psy-
chophysics, that ultimately led to signal detection theory.
The chart of Figure 1, however, does not imply that those
who pioneered the research leading to signal detection
theory necessarily knew of Fechner's contributions: The
chart is in terms of commonalities of interest rather than
of direct historical interaction. It must be understood that
Fechner himself saw inner psychophysics as the central
core of his theory and said that a psychophysical law
applied particularly to inner psychophysics; it would not
be expected to be applicable to outer psychophysics and
to fit the data perfectly, because there are many sources of
variability determining the data of outer psychophysics.
We can now trace the development of these two strands,
reviewing first the nineteenth-century evidence concern-
ing alternative psychophysical laws.
1.3. The line of research leading
from "outer psychophysics"
1.3.1. Retaining the logarithm: Delboeuf's equation.
Apart from the logical rigour of its derivation, the most
powerful evidence in favour of a logarithmic law relating
sensation intensity to physical intensity was believed to
be that obtained from a series of experiments carried out
by J. R. L. Delboeuf (1831-1896), a scholar associated
with the University of Liege in Belgium. Delboeuf wrote
a monograph of 115 pages in 1873 in which he first
reviewed Fechner's derivation of Fechner's Law and then
he also reviewed a modification of the law that had been
made by Helmholtz (1856-66/1962). According to Helm-
holtz, any new sensation of brightness is added to a
"natural light of the retina," and this is in turn a function of
the total ambient light in a perceptual situation. For
example, if a contrast is set up between a light object and
another object in the surround of nighttime illumination,
the contrast looks less striking if the two objects are then
viewed under daytime illumination. Helmholtz sug-
gested a modified version of Fechner's fundamental for-
mula, one which made an important qualitative change.
In our notation, Helmholtz suggested that
I')]
(5)
where /' is a function of the natural light of the retina and
A is a c6nstant. Possibly inspired by Helmholtz's intro-
duction of a factor indicative of a resting state of activity in
the nervous system, Delboeuf went on to derive an
equation of his own; in so doing he extensively discussed
such factors as the degree of adaptation or fatigue in
receptor systems, but the end result was that Delboeuf's
equation, in its simplest form, was:
Murray: History of psychophysics
S = log [(c + l)lc] (6)
where c is a constant, to be evaluated by experiment, that
reflects the ongoing state of activity in the nervous system
at the time the stimulation measured by Z is presented.
One important fact to note about Delboeuf's formulation
is that, like Fechner's, it is logarithmic. Another is that
Delboeuf started using a new method for estimating
sensation strength, a method based on estimating what
stimulus intensity should be chosen so that it yielded a
sensation that looked halfway between two other sensa-
tions. The originator of the idea that sensation strength
could be measured by finding a sensation midway be-
tween two others was J. A. F. Plateau (1801-1883) of
Brussels; we shall describe his findings below. Delboeuf,
however, only contacted Plateau about his own experi-
ments after he had started them; they were designed
explicitly to test Fechner's theory.
Delboeuf, then, set out to examine the validity of
Equation 6 by carrying out a systematic series of experi-
ments in which he obtained any desired grey by revolving
a black disc that contained a white sector of a given
angular width: The greater the area of white, the lighter
the resulting grey. His visual display was presented
against a black background, with the whole set-up being
viewed in a variety of daylight illuminations or by candle-
light. The display was composed of three concentric
rings, a dark outer ring, an intermediate middle ring, and
a light inner ring (see Figure 2). In all his experiments,
the greyness of the outer and middle rings was prear-
ranged and the subject had to set the greyness of the inner
ring so that the middle ring looked intermediate between
the outer (dark) and the inner one (light, set by the
subject). If the three greynesses thus represented are
labelled as 7X (outer darker ring), 72 (middle ring), and 73
(inner ring), each with its corresponding sensation S j , S2,
and S3, then from Equation 6, we surmise that
S, = log[(c + /,)/<:]; S2 = log [(c + Qlc\,
S3 = log [(c + !3)/c)] (7)
and moreover that
Sz-S^Sz- S2 (8)
From these identities we can express c in terms of Iu Z2
and Z3; in fact, Delboeuf deliberately picked values oflu
Z2, and Z3 so that Z2 looked intermediate between lx and Z3
for a particular placement of a candle, and he solved for c;
he then used this value of c to calculate what Z3 ought to
be, given various other values of 7X and Z2. This calculation
of an estimated Z3 was based on the following relationship,
derived from Equations 7 and 8:
73 = (/| - c/, + 2cZ2)/(c + /,) (9)
The question of course was whether the estimated Z3 as
calculated in this way matched the Z3 that was actually set
by the subjects. In Delboeuf's first table, 13 different
values of Zj and Z2 are investigated and shown along with
the subject's settings of Z3; each setting of Z3is the average
of five settings and Delboeuf stressed that the subject
could often take a long time (minutes) in choosing Z3.
Moreover, the overall level of daylight illumination, gen-
erally resulting from a cloudy sky, could not be con-
trolled. Delboeuf quoted two measures of central ten-
dency, the halfway mark between the highest and lowest
scores of the five settings of Z3, and the arithmetic mean of
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 119
Murray: History of psychophysics
BLACK SURROUND

BLACK SURROUND

ooo
Background i Background)
preset preset

1 preset \ / variable \

One preset darkest One preset lighter One variable

(a) (b) (c)

Figure 2. The displays used by Delboeuf (1873) and Lehmann (1886). (a): In Delboeuf s display the subject must
set 73 so that 72 looks intermediate in lightness between lx and 73. (b): Lehmann's first display shows that the variable
disc must be adjusted by the subject to look intermediate between the darkest and the lighter disc. The effect of the
position of the variable disc in the row was systematically explored, (c): In Lehmann's second display the variable
disc had to be set at a lightness r that matched the lightness of the test disc i. Any "mismatch" would indicate a
contrast effect induced in the appearance of the variable disc by the background /.
the five scores. Both measures were very similar. Del- of 73. Delboeuf was content to point to the apparent
boeuf also calculated the estimated 73 by using two values goodness of fit; Fechner (1877, pp. 178-87) actually
of c, an upper and lower estimate, but again the two attempted to measure the goodness offitby calculating
values of 73 obtained in this way were very similar. Table 1 the squares of the deviations between the estimated and
is a partial reproduction of Delboeuf's first table. Even the obtained values of 73 for a number of Delboeuf's
though these results are only for a single subject, the tables, finding the bestfitto hold for the particular table
historical importance of this table justifies its presentation shown here. Fechner also recalculated the values of 73,
here. however, making the assumption that c could approach 0;
It can be seen that there is an extraordinarily good at very small values of c, he claimed, Equation 6 came to
match between the predicted 13 and the obtained settings approximate Fechner's own Equation 4. Fechner found
that this amendment altered the goodness of fit only
Table 1. Results of Delboeuf s bisection experiment slightly, and in later writings Delboeuf seems to have felt
that Equation 6 was not necessarily much better than
Greyness Greyness Predicted Median Mean Equation 4. The fact remains, however, that in this study
of outer of middle value of 73 setting setting Delboeuf presented evidence he believed supported his
ring (7j) ring (/g) for c = .5 of7 3 of7 3 modification of Fechner's formula.3
9 47 237 232 237.6 1.3.2. Sensation differences versus sensation ratios.
13 27 55.5 55.5 54.4
Delboeuf also indicated that, apart from the degree of
13 36 98.3 105 98.8 goodness of fit, there was a further way in which his
13 41 127 128 129.2
formula could be said to be superior to that of Fechner.
13 56 236 274 (?) 247.8
The reader at this point should consult Table 2, which
21 60 169.7 167 163.4
represents a proof by numerical demonstration that, if
21 64 193 198.5 200
Fechner is right, when a given difference between sensa-
22 36 58.7 57.5 57.6
tions (call this a "contrast") is set up, then brightening the
22 51 117.4 115 119.8
overall illumination of the display associated with that
22 58 151.6 151 152.2
contrast should not change the apparent degree of the
43 64 97.4 97.5 97.4
contrast. If Delboeuf's psychophysical law is correct, it
43 72 119.5 131.5 130
predicts that if we vary the overall illumination by bright-
43 87 175.5 171 176.8
ening it, the contrast that was set up before the change in
illumination will look different following the change. In
Note: An abbreviated version of Table 1 from Delboeuf (1873). Table 2 we see that if we brighten the overall illumination
73 is the greyness allotted to the setting of the inner ring (see by a factor of 5, the difference between the two resulting
Figure 2), adjusted so that it makes 72 look intermediate be- sensations increases, which suggests that the contrast
tween 73 and Iv The values of 7 t and 72 (set by the experi- should be enhanced. It follows that darkening the display
menter) are shown in columns 1 and 2; the median and mean should reduce the contrast. On the other hand, if we
settings of 73 (set by the subject) are shown in columns 4 and 5; brighten the overall illumination by a factor of 5, the ratio
and Delboeuf s predicted setting of 73, based on Equation 8, is between the two resulting sensations decreases, which
shown in the third column. The units represent the angular suggests that the appearance of contrast should be less-
width subtended by a sector of white in an otherwise black ened. Modern research, discussed in section 2.2 below,
ring: The lower the number, the darker the appearance of the hasfindingsof relevance to an evaluation of this unusuai
ring. outcome of Delboeuf's Law, but we note here that Del-

120 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1


Table 2. A numerical demonstration of predictions generated by Fechner, Plateau, and Delboeuf on whether a contrast set up
under one illumination (in which stimulus ratios are 2:1) will change in appearance if the overall illumination is brightened
(making the stimulus ratios 10:5 respectively)
Fechner
s=yiiog/)
Take logs
to base 10
Stimulus
Stimuli intensity I Sensation
A 2 log 2 = 0.301
B 1 log 1 = 0.000
0.301 diff.
9.427 ratio"
C 10 log 10 = 1.000
D 5 log 5 = 0.699
0.301 diff.
1.431 ratio
If sensation differences are
judged, a contrast under
brighter illumination will look: Unchanged
If sensation ratios are judged,
a contrast under brighter
illumination will look: Reduced
"using log 1.01
boeuf himself made a simple demonstration that a given
contrast could be reduced if overall illumination was
darkened, in line with the prediction from his law on the
basis of sensation differences. On a white disc he drew a
portion of a narrow black sector. When the disc was
rotated, a grey zone appeared on the white surface of the
disc. If this zone was just visible by the light of a candle
placed nearby, it became invisible if the candle was
moved away from the display to a considerable distance:
To make the zone visible again, the angular width of the
sector of which a portion was drawn had to be increased.
Two points of general interest may be noted here.
First, consider the special case of two stimuli A and B
where B is just noticeably different from A. According to
Fechner's Law, you should be able to darken or brighten
the scene, and the contrast between A and B will be such
that B continues to look just noticeably different from A.
Fechner's own demonstration was to ask the observer to
pick out a wisp of cloud that was just noticeably different
from the background of the sky and then look at the cloud
through a dark glass. He predicted that the cloud would
continue to be just noticeably different from the sky. If
Delboeuf was right, there should be a limit to how far this
is true; there should be a particular shade of darkness in
the glass where the contrast is so reduced that the cloud
becomes indiscriminable from the sky. Second, it is
possible that in contrast situations white areas have a
darkening effect on dark areas (so that increasing illu-
mination could make dark areas even blacker), whereas
dark areas may have no effect on white areas. This matter
will be taken up later, but we would note that this kind of
asymmetrical effect would be consistent with Delboeuf's
results.
Murray: History of psychophysics
Plateau Delboeuf
S = log[(c + l)lc]
s =m Let c = .5
Let n = .5
Sensation Sensation
2 5 = 1.414 log (2.5/.5) = 0 . 6 9 9
15 = 1.000 log (1.5/.5) == 0.477
0.414 diff. 0.222 diff.
1.414 ratio 1.465 ratio
10 5 = 3 . 1 6 2 log (10.5/.5) = 1.322
5 5 = 2.236 log ( 5.5/.5) = 1.041
0.926 diff. 0.281 diff.
1.414 ratio 1.269 ratio
Enhanced Enhanced
Unchanged Reduced
1.3.3. Rejecting the logarithm: Plateau's equation. As
mentioned earlier, Plateau (1872a) devised the method
of equisection, now known as the method of equal-
appearing intervals. He did it extremely informally:
When he asked eight painters to mix a grey that appeared
equidistant between white and black he found that the
greys thus produced were very similar. He also noted that
an etching, consisting of strong contrasts of black and
white, seemed perceptually identical in terms of contrast
whether viewed under daylight, strong sunlight, or can-
dlelight. According to our Table 2, this would be expected
if apparent differences between adjacent areas are what
remain constant as overall illumination changes, and
Plateau (1872a, p. 9) explicitly stated: "Fechner's formula
has the consequence that, if the overall illumination
varies, it is differences between sensations that remain
constant." Nevertheless, Plateau felt that it was not nec-
essarily differences between adjacent areas that we would
expect to remain constant, but ratios, or relations, be-
tween adjacent areas; he went on to remark: "It appears to
be more reasonable, that if we are to understand the
persistence of the general appearance of the etching, to
accept a priori the constancy of the ratios, not the differ-
ences, between sensations. "As Table 2 shows, if ratios are
to be preserved under different illuminations then the
appropriate psychophysical law is a power function.
Plateau proved this algebraically; Brentano (1874/1973)
commented that he felt that equal stimulus ratios implied
equal sensation ratios; and it was Fechner himself (1877,
p. 24) who proved that a modification of the fundamental
formula (Equation 2) would lead to a power function.4 The
modification consisted of assuming that it was not AS
alone that was incremented if the starting sensation /
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 121
Murray: History of psychophysics
were increased by A/, but the relation between AS and
the sensation S aroused by I
&S/S = C(M/I) (10)
Solving this equation by the same chain of argument that
led to Equation 4, the logarithmic law, leads to the power
law
S = Klc (11)
where K is a constant. 5
Yet despite this intense interest in the possibility of a
power law, it was Plate' a (1872b) who, after reviewing
Delboeuf's paper, came to the conclusion that Delboeuf's
equation was superior both to Fechner's equation and to a
power law when it came to predicting the data. More-
over, Delboeuf's equation predicts that contrasts will
change in appearance if overall illumination changes,
whereas Plateau's equation had predicted that if apparent
ratios were preserved they would not. Hence Plateau
(1872b) retracted his views in the following words:
As for my formula, from the mere fact that it differs
from that of M. Delboeuf, it is clear that it is inexact. It
rests on the hypothesis that the level of illumination
influences neither the ratios of sensations nor the
equality of apparent contrast; but it can be easily
demonstrated using M. Delboeuf's formula as a start-
ing point, that the ratios of the sensations do depend on
the level of illumination, and it has been seen that the
equality of the contrasts also depends on this, though in
a less pronounced fashion, (p. 261)
At the risk of interrupting the historical flow, we also
note a problem here that did not receive sufficient atten-
tion in earlier years, at least until the work of Torgerson
(1961); in Plateau's bisection task, the subject is not asked
to choose 72 so that S2/S1 = S3/S2, he is asked to choose 72
so that Sz S1 = S3 S2. The question may be raised
whether a subject asked to do the former task - preserv-
ing ratios - would produce contrasts that are preserved
under different illuminations, a finding that would favour
a power law. Over the past two decades, however, it has
become increasingly apparent that task performance is a
complicated function of whether subjects are asked to
judge differences or ratios. Put overbriefly: Torgerson
(1961) suggested that subjects made all their judgments
on the basis of a sense of "difference' or "distinctiveness"
and that it might be misleading to assert that a given
contrast is arrived at by a subjective judgment of ratio.
Birnbaum and Veit (1974) showed that subjects could say,
by giving a number, what the difference was between two
weights (one in each hand) and what the ratio was be-
tween the weights. However, following a complicated
literature, Birnbaum (1980) came to maintain that a single
mental operation underlay both difference judgments
and ratio judgments and that this single operation re-
flected a subtractive procedure, so that a given difference
sensation Sab was equivalent to a sensation difference (Sa
Sj,). There is a current of modern opinion to the effect
that when a person sets up a "contrast" between two
stimuli, A and B, he is judging their sensation difference
rather than their sensation ratio. To be weighed against
this claim are (a) evidence that some subjects sponta-
neously give ratio judgments even when the task is
devised to yield difference judgments (Popper et al. 1986;
see also Schneider 1980a; 1980b); (b) evidence that when
122 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
subjects are asked to judge which is the larger of two
differences in loudness (a - b) or (c - d), a scale can be
constructed in which estimated loudness is not a log-
arithmic but a power function of stimulus intensity
(Schneider 1980a); and (c) evidence that some sensory
dimensions may be associated with relatively easy differ-
ence judgments (such as loudness, as just noted), whereas
on other sensory dimensions subjects more naturally
make ratio judgments. Among the latter dimensions are
length (Parker et al. 1975), area, volume (Schneider &
Bissett 1988), and possibly pitch (Schneider et al. 1982).
Schneider and Bissett (1988) have shown that this distinc-
tion may be unrelated to the prothetic/metathetic dichot-
omy; instead, it seems that dimensions such as length,
area, or volume, along which stimuli are easily "decom-
posed" into a number of smaller perceptual units, are
more likely to be associated with ratio judgments,
whereas dimensions such as loudness or heaviness (see
Rule et al. 1981), where such decomposition is subjec-
tively unfeasible, are more likely to be judged in terms of
difference judgments.
Returning to the historical argument: Despite Plateau's
retraction, the power law did not vanish. Even Fechner
(1887), in his last article, admitted that if the fundamental
formula of Equation 10 were valid then a power law could
be a reasonable contender for the psychophysical law. He
claimed, however, that the logarithmic law worked better
in practice; moreover, he stuck tenaciously to the opinion
that the power law suffered from a serious drawback,
namely, its assertion that if / were nonzero, the subject
would necessarily have a sensation S of some strength,
since the constants K and C in Equation 11 were both
positive. For Fechner, building on the inner psycho-
physics described earlier it was reasonable to suppose
that an intensity 7 could be positive yet unable to arouse a
sensation because it elicited excitation below the thresh-
old required for a conscious experience. As noted above,
his opponents misinterpreted his use of negative signs to
refer to such subthreshold excitations as implying that
subjects had "negative sensations," an incomprehensible
concept; Fechner (1887) wearily remarked that "it is
incontestably to be desired that the controversy over
negative sensations should come to an end once and for
all; but if my experience up to now is my guide, even my
ghost will have no peace because of it" (p. 224). For
Fechner, it was a fault, not a virtue, of a power law that it
implied there could be no subthreshold excitation arising
from a positive intensity. Modern writers on psycho-
physics, however, have suggested several ways in which
subthreshold excitation arising from a positive intensity
can be reconciled with a power law (see e.g., Atkinson
1982; Dawson & Miller 1978).
1.3.4. Evidence for the power law based on the inequality
of difference sensations. Another reason the power law
continued to hold sway arose initially from an observation
by Hering (1875) attacking the view that, throughout a
range of/ values, AS is constant. According to Hering, a
difference sensation arising from an 7X (say a weight of 2
units) and an 72 (say a weight of 1 unit) is not perceived to
be the same difference sensation as arises from two other
weights (say, a weight of 10 units and a weight of 5 units)
that bear the same ratio as 7j to 72. Hering actually did the
experiment, placing weights of 500 and 1,000 grams on

one hand and weights of 1,000 and 2,000 grains on the
other; the latter difference, he claimed, seemed larger.
Fechner simply denied that the experiment was intro-
spectively valid: While the weights on the second hand
were certainly heavier in toto, the task of judging differ-
ences was too difficult using this paradigm (in fact Fech-
ner was always worried about the difficulty of evaluating
either AS or Sl S2, where the two stimuli involved were
separate in space or time, a worry shared by Krueger
1989r, sect. 2.3).
Nevertheless, Stevens (1957) has made it a cornerstone
of his belief in a power law that if a jnd (S1 S2) is
established for a pair of low intensities Ix and 72, the jnd
(S3 S4) that is established for a pair of higher intensities
73 and 74 on the same sensory dimension is not necessarily
equal in apparent "size" to (Sv - S2). This assertion can be
validated by establishing (Sl - S2) and (S3 - S4) by a jnd
method or alternatively showing that the difference be-
tween sensory magnitude estimates of Ix and 72 is less
than the difference between sensory magnitude estimates
of/3 and 74. Stevens proved that if a power law holds, and
if Weber's Law is true, then the subjective size of the jnd
increases as an exponential function of the number ofjnds
by which the stimulus is above threshold (Stevens 1957,
p. 172). Later, Ekinan (1956; 1959) formulated "Ekman's
Law," which stated that the psychological size of the jnd is
a linear function of the sensation magnitude it is based on;
Teghtsoonian (1971; 1973) amended Ekman's Law to take
account of the range of stimulus magnitudes used in an
experiment - subsequent experimentation, such as that
of Marks et al. (1986) on cross-modality difference estima-
tion, has shown the influence of the range of stimulus
magnitudes on subjects' responding. The persuasiveness
of Ekman's argument was well illustrated by Gescheider
(1976, p. 160), who gave a proof by numerical demonstra-
tion showing that if both Weber's and Ekman's laws are
true then a power function of the type favored by Stevens
would be the resulting psychophysical law connecting S
with /. Special note should be made of a claim concerning
loudness by Parker and Schneider (1980) according to
which, if Weber's Law is false for a sensory dimension, it
is possible to have a psychophysical law that is a power
function as well as subjective jnds that do appear constant
in size across a range of loudness, as Fechner postulated.
That Weber's Law might not hold for the loudness of
pulse sinusoid tones is suggested by data obtained by
Jesteadt et al. (1977). In this study, log (A7/7) did not
remain constant, but decreased linearly, as a function of
the stimulus intensity measured in decibels SL. The
tones varied in frequency between 200 and 8,000 Hz.
For the moment we rest content with pointing to the
enormous popularity of a power law in twentieth-century
research following Stevens's espousal of it. As an example
of its general applicability, we can refer to Table 5.1 in
Gescheider (1976, p. 129), which shows that in 33 experi-
ments dealing with various stimulus dimensions, power
laws could be fit to the data relating S to 7 with exponents
varying from .33 (brightness as a function of the lumi-
nance of a 5 target in the dark) to 1.0 (apparent length of a
projected line) to 3.5 (the apparent "shock" value of an
electric current passed through the fingers). The fact that
some of these exponents can be greater than one illus-
trates the generality of the power law. A logarithmic law is
associated with a negatively accelerated curve once the
Murray: History of psychophysics
point where 7/70 equals e (2.71828) is passed (see Fechner
1860/1964, vol. 2, p. 49), a feature that is often over-
looked; a logarithmic law cannot handle a positively
accelerated curve such as is apparently obtained with
rated heaviness, tactile roughness, and subjective
warmth associated with a metal object touching the arm.
The exponent of 3.5 quoted above for electric shock has to
be balanced against a claim of Swartz (1953): For three
different ranges of current applied to the tooth pulp to
yield sensations of pain, if the subject set a variable
resistor to yield sensations half-way between the thresh-
old level and the pain tolerance level, the resulting graph
of bisection points against actual fractions of the range of
stimuli was linear. Power laws can certainly be fitted to
scales of "sensation strength" based on magnitude esti-
mates and cross-modal matching, and Stevens (1970)
claimed they could also be fitted to neurelectric data. In
fact, here Stevens seemed to argue that the nonlinear
transformation of an increase in stimulus intensity into an
increase in sensation strength took place in the receptor
rather than in the brain, as Fechner had claimed. The
present situation, to judge by the responses to Krueger's
attempt to show that a modified power law can also be
applied to summated jnd scales and category judgments,
is that the controversy sparked in the nineteenth century
by Plateau is still with us. In fact, we may say that it is the
major issue that arose from Fechner's outer psycho-
physics.
Figure 1 suggests that modern psychophysics stems
from two main lines, the line just described, leading from
outer psychophysics, and a second line arising from Fech-
ner's definition of inner psychophysics. We now turn to
this second line of research.
1.4. The line of research leading
from "inner psychophysics"
As noted, Fechner assumed that the excitation arising
from a stimulus was added to a resting level of excitation
and indeed that this excitation might have ensuing excita-
tions that had to cross several "thresholds" before the final
threshold leading to a sensation, a conscious experience,
was breached. We also noted Helmholtz's assumption
that at the level of the measurement of sensation whether
or not an increment in visual sensation was detected
depended upon the increment's being large enough to
exceed "natural light in the retina," whereas Delboeuf
assumed that a variable he called c had to be taken into
account in the psychophysical law. According to Krueger
(personal communication) it may even be possible to
reconcile Delboeuf's model with Fechner's if a nonzero
(but subthreshold) value of 7 could induce a nonzero S that
was nevertheless not reflected in conscious experience.
Krueger has also pointed out that according to Fechner's
Law, when 7 = I o , S must equal zero, yet it is at that point,
the absolute threshold, that'S becomes consciously per-
ceptible. Fechner wrote that this was proof that a thresh-
old must exist, for at a finite value of 7, namely 70,
sensation value S must be zero (Fechner 1860/1964, vol.
2, p. 14). This does not rest easily, however, with those
who would make the sensation at the absolute threshold a
conscious experience.
The notion that a stimulus is never processed in isola-
tion, but only against a background, holds both for physi-
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 123
Murray: History of psychophysics
ological discussions of sensory excitation and psychologi-
cal discussions of increments in sensations. Later we will
examine the relationship between these two types of
discussion in more detail, but as a historical note we
would remark that Delboeuf s psychophysical equation
(Equation 5) was derived from an assumption about neur-
electric events. The same can be said of an argument
produced by our next figure, L. M. Solomons, who in
1900 offered an explanation of why Weber's Law took its
particular form in terms of a discussion of random activity
in the brain. Solomons assumed that any brain activity
underlying a sensation was not fixed but variable, and that
sensation in turn was itself marked by variability. He
presumed that the degree of variability in brain activity B
associated with stimulation from a physical activity /
varied between B AB and B + AB, with AB itself being
a function of B. For an increment in intensity 7 + A7 to
yield a sensation just noticeably different from that
aroused by 7, 7 + A7 had to set up brain activity not merely
exceeding B but also exceeding B + AB. Since AB was
held to be a constant fraction of B, AZ would have to be a
constant fraction of 7 if / + A/ were to be discriminated
from 7. This, of course, is Weber's Law. But Solomons's
name is mentioned not merely to add to the list of
theorists who argued that stimulations have to exceed
resting levels, but because he was one of the first to stress
that variability in the sensory system is associated with
variability in sensation and that any theory of psycho-
physics has to take variability not merely as a source of
error but as a given quantity to be evaluated. We have
already noted that Fechner measured variability. The
idea that variability is a useful measure per se in the
estimation of sensation strength was criticised by Stevens
(1957).
1.4.1. The psychometric function. The point of contact
between the line of research concerned with the psycho-
physical law (shown in Figure 1) and the line of research
concerned with variability and "noise" lies in what is
known as the psychometric function. In establishing just
noticeable differences, the method of right and wrong
cases involves presenting two stimuli, A and B (with A and
B being close to each other on a physical continuum of
weight, length, intensity, etc.) and having the subject
decide whether A > B, "yes" or "no." If we plot the
percentage of yes judgments (p) against the size of the
physical difference (A - B), we typically find an approx-
imately ogival curve rising from p = 0 for small values of
(A - B) to p = 1.00 at large values. The term "psycho-
metric function" was given common currency by Urban
(1910), but prior to Urban's time a number of authors had
claimed that the mere fact that the rise in the curve from p
= 0 to p = 1 is not discontinuous, but continuous,
suggested (a) that the definition of what constituted a
threshold value on the A B continuum was necessarily
arbitrary and (b) that the continuous nature of the curve
indicated that in the nervous system itself there was not
necessarily a strong continuity between "no sensation"
and "some sensation." As Corso (1963) outlines in detail,
suspicions as to the value of the concept of a neurelectric
threshold were expressed by G. E. Muller(1878), Jastrow
(1888), and Cattell (1893); Jastrow wrote:
The conception [of the threshold] grew directly out
of the method of the just observable difference; in fact
124 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
this difference has by some been taken to be the
differential threshold. What is most usually denoted by
the threshold is the smallest difference that can be
perceived. It is the threshold of conscious-
ness . . . either there is a threshold - be it a point or a
more or less variable line - below which is homoge-
neous unconsciousness; or from the region in which the
sensed difference has its maximum of clearness down to
the point where it utterly vanishes because this differ-
ence between the stimuli vanishes, there is a contin-
uous series of intermediate degrees of clearness, and
there is no point on the curve with characteristics
peculiar to itself, no threshold in any true sense, (pp.
276-77)
But this skepticism did not stop others from assuming that
the threshold could be defined as a particular point in the
psychometric function: Urban (1910) gave details for
estimating the midpoint of such a function from raw data,
and Boring (1917) gave an account of the psychometric
function showing how it could arise if there were random
fluctuations in responding caused by a number of factors
that could not be experimentally controlled.
The theory that the psychometric function was an
artefact of chance variables disguising the true sensation
of a difference between A and B was called the phi-gamma
hypothesis by Urban and Boring.6 In Figure 1 we have
indicated that Fechner's inner psychophysics formed a
historical backdrop to research that ultimately culmi-
nated in signal detection theory only because Fechner's
inner psychophysics stressed that sensations were the
outcome of stimulus-produced oscillations superimposed
on resting-level oscillations; it was the doubt raised by
observers of the psychometric function as to whether a
sensation arose discontinuously from these superimposed
oscillations that instigated subsequent research on vari-
ability, and thence signal detection theory. By the time
signal detection theory had arrived, doubts were being
expressed as to the existence of a discontinuous "thresh-
old," and the line in Figure 1 connecting Fechner's inner
psychophysics with signal detection theory should be
interpreted as indicating a common thread (signal against
noise) but also a contradiction (concerning whether or not
there was a discontinuous threshold). By the time signal
detection theory appeared in the context of psycho-
physics (Swets et al. 1961) the problem of "variability" was
being related to the problem of "noise" of various kinds:
(1) noise intrinsic to the stimulus, as exemplified by the
quantum nature of light; (2) noise as a physical back-
ground to the stimulus; (3) noise in the sensory receptor;
(4) noise in the brain; and (5) noise in the sense that
subjects' responses were variable because of biases,
order-of-presentation effects, fatigue, and so on (but
see, e.g., Wertheimer 1953, for the claim that "order-
of-presentation effects should not be considered as ran-
dom").
A continuous psychometric function might be influ-
enced by noise of any of the above types, as Urban and
Boring had suggested; but it was also claimed, partic-
ularly by Stevens et al. (1941) and Corso (1967, pp. 423-
45), that a continuously increasing psychometric func-
tion, particularly if it was linear as opposed to ogival, was
consistent with the hypothesis that the sensory receptor
responded in a quantal fashion, and that if a receptor was
already firing to a given stimulus intensity, there were

ranges of stimulus intensities over which no changes in
firing would take place until a particular value of stimulus
intensity (below or above the original) had been reached.
This limit on receptor sensitivity can be described in
terms of noise of type (3), and Cornsweet (1970, Ch. 4) has
a particularly detailed analysis of receptor noise in the
context of neural quantum theory. Thresholds for firing in
receptors were later distinguished from thresholds con-
tingent upon brain and response variables ("observer"
thresholds), and Luce (1963) and Krantz (1969) showed
how, if observer thresholds existed, it could be predicted
that curves of receiver-operating characteristics could be
rising discontinuously as opposed to continuously.
Discontinuities in sensory functioning are also illus-
trated by the phenomenon of "categorical perception," in
which only a few stimulus values within a range of possi-
ble stimulus values along a dimension can be discrimi-
nated; a variety of theories as to why this should be the
case are discussed in the book Categorical perception
edited by Hamad (1987): A particular problem raised by
Macmillan (1987), Massaro (1987), and Pastore (1987)
concerns the nature of the relationship between the
discrimination of a difference between two stimuli and
the facility with which the two stimuli can be assigned to a
category. There is agreement that context effects, anchor
effects, and memory demands affect both tasks, but par-
ticularly the latter, whereas within a "category" discrimi-
nations between individual stimuli are often possible
given appropriate psychophysical tasks. Because the as-
signment of a category reflects a dimensional process,
both Macmillan (1987) and Rozsypol et al. (1985) have
suggested that signal detection theory can be applied to
categorical perception tasks. Wilson (1987) has argued
that the various contextual effects that influence a deci-
sion concerning which category a stimulus belongs to can
be modelled by adaptation level theory.
From a historical point of view, the link between the
phi-gamma hypothesis of the 1920s (with its generalized
treatment of variability) and the signal detection theory of
the 1960s (with its analytical treatment of various kinds of
noise) was provided by Thurstone (1927a; 1927b; 1927c;
1927d). Instead of trying to establish a "threshold" from
the psychometric function, Thurstone argued that each
stimulus, A and B, set up a range of internal states (a1,
o 2 . . . a j a n d ^ j , b2 bn). Each of these states had a
possible frequency, so that the states (alt a2 an), for
example, could be taken as normally distributed with a
particular state a representing a mean subjective state
associated with stimulus A, and similarly, another state b
representing a mean subjective state associated with
stimulus B. Using what is now known as the method of
paired comparisons, Thurstone established the "law of
comparative judgment" which took various forms de-
pending on the degree to which A and B were considered
independent: For complete independence of A and B,
Pnb = N(a-b) (12)
where p a b is the probability that A is judged greater than
B and N is the cumulative normal distribution function.
This is a rephrasing of the equation for the psychometric
function in which a and b represent mean sensory states
out of the many that can be aroused by each of A and B.
Incidentally, Thurstone's work started a trend in which
deterministic mathematical models of psychological pro-
Murray: History of psychophysics
cesses were often modified to make them probabilistic
(see Falmagne 1986, for several examples of this trend in
psychophysical theory).
Thurstone's model can be diagramed by showing a
range of sensory states along an axis and drawing two
normal distributions on top of the axis, one representing
the range of states excited by stimulus A (with its mean at
a) and the other representing the range of states excited
by stimulus B (with its mean at b). This diagram is shown
in Gigerenzer and Murray (1987, p. 37) but they go on to
argue that Thurstone for some reason did not add a
decision rule, that is, a rule determining which response,
yes or no, will be given if a particular state is experienced.
The addition of such a rule was provided by the signal
detection theorists in the 1950s; Gigerenzer and Murray
show in detail how the typical signal detection model
showing two normal distributions of sensory states, one
associated with noise and the other with signal-plus-
noise, along with a movable criterion, is an analogue of
the Neyman-Pearson model of inference-making con-
cerning two hypotheses. J. A. Swets (personal communi-
cation) has suggested that the main innovation of the
researchers on signal detection theory was to postulate a
movable criterion. In modern terms, this means that
when we are attempting to establish thresholds, the task
involves a certain level of "cognitive penetrability," as
Pylyshyn (1984) expresses it [see also Pylyshyn: "Com-
putation and Cognition: Issues in the Foundations of
Cognitive Science" BBS 3(1) 1980].
1.5. Some current Issues
It is not my purpose here to review the enormous amount
of modern work either on signal detection theory (see
e.g., Falmagne 1985, Chapter 10; Macmillan & Creelman
1991; McNicol 1972; Swets et al. 1961) or on the psycho-
physical law (e.g., Falmagne 1985, Chapters 3, 14;
Krueger 1989a; 1989r). If we approach the signal detec-
tion model as a model of brain activity, the following
matters remain open:
1.5.1. On signal detection theory. It has been frequently
claimed (see e.g., Green & Swets 1966) that as signal
strength increases, so does the variability of the underly-
ing signal-plus-noise distribution. If so, this would be
consistent with Solomons's suggestion that the effect of
range of brain states increases as stimulus intensity in-
creases. Yet we have found little experimental work using
signal detection theory, and a variety of signal strengths,
apart from that of Markowitz and Swets (1967), devoted to
establishing whether this claim is true. Moreover, Swets
(1986a; 1986b) examined both theoretical and empirical
receiver operating characteristic (ROC) curves derived
from a variety of indices relating hit rates to false alarm
rates and came to the conclusion that there was no
common explanation for cases of unequal variances of
signal and noise distributions as estimated from the slopes
of empirical ROC curves. Indeed, one of the studies
surveyed related to recognition memory, where it was
shown that the linear ROC curves relating to memory for
words studied once was nearer the chance line than was
that relating to words studied twice; but the slopes of the
two curves were both about 1.00 (Egan 1958). Neverthe-
less, research by Tolhurst et al. (1983) indicates that an
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 125
Murray: History of psychophysics
increase in visual signal strength causes not only an
increase in the mean number of impulses in single cells in
the visual cortex but also an increase in the variance
associated with that number. Furthermore, Thurstone's
Case V assumes that if we have two mean states of
sensation strength a and b, with b > a, the distributions
associated with b and a are supposed to have equal
variability. But if the variability associated with b is
greater than that associated with a, and this variability is
held to be a function of sensation magnitude, then scales
obtained from the Thurstonian paired comparisons
method would be linearly related to scales obtained from
ratio methods.
Other questions on which considerable literatures now
exist concern how the criterion location is related to a
variety of stimulus parameters, including the proportion
of signals in the experiment (Parks 1966), the effect of a
signal presented prior to a critical signal (Treisman
1964b), and the effect of varying the signal-to-noise ratio
while the response bias is kept constant (Dusoir 1975;
1983; see also Curry et al. 1977). Much of this empirical
work on signal detection has concerned signals varying
along one dimension only, but the theory can be extended
to multidimensional signals; according to Macmillan and
Creelman (1991, p. 240), Tanner (1956) first proposed
measuring how a stimulus of one frequency, Su could be
discriminated from noise, measuring how a stimulus of a
different frequency, S2, could be discriminated from
noise, and then measuring how discriminable Sj was from
S2. These questions have all arisen within the context of
the signal detection model, however; recent attempts to
relate signal detection parameters to those obtained from
other models have focused on relationships with Luce's
choice theory rather than with Thurstone's theory (see
e.g., Swets 1986a; Macmillan & Creelman 1991).
1.5.2. On the psychophysical law. The question has been
raised whether experimental data can be used to decide
between a logarithmic and a power law of psychophysical
functioning. Krueger (1989a; 1989r) has pointed out that if
a power law has a very small exponent, its predictions
cannot be distinguished from those of a logarithmic law
(see also footnote 3).
As Krueger's 1989 article and its accompanying com-
mentaries indicate, at least some members of the psycho-
physics community would like to integrate findings from
scales based on jnds, bisection methods, category judg-
ments, and magnitude estimates within one broad theo-
retical framework (Krueger himself has suggested a mod-
ified power law as one possible framework). As a side
issue, the question has been raised as to the nature of the
internal "scale" representing the natural numbers, a scale
presumably used in the assignment of category judg-
ments and magnitude estimates (Baird & Noma 1975).
The nineteenth-century dream of starting with as-
sumptions about neurelectric phenomena and from these
deriving, by as rigorous a deductive chain as possible, a
psychophysical law, or failing this, an account of receiver
operating characteristics, is very much alive in the writ-
ings of D. Laming (1985; 1986; see also Laming 1989;
1991). One of Laming's discoveries is that tasks involving
judgments comparing two adjacent stimuli (the "detec-
tion of differences") can yield estimates of differential
thresholds or ROC curves that differ from those obtained
126 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
from tasks when the subject must detect an increment
added to a stimulus ("increment detection"). This finding
would hearten Fechner, who always claimed that the
measurement of differential thresholds was extremely
difficult because of the inability to control all aspects of
the display.
1.5.3. Von Kries. We suggest that the scheme shown in
Figure 1 can facilitate thinking about psychophysics in
general, but before turning to the second aspect of our
"perspective," we must briefly mention that Fechner's
work was criticised by von Kries (1882) on the grounds
that Fechner's whole endeavour rested on quicksand:
Sensations are not the sort of thing that can be measured
in the way that length or time can. Von Kries claimed that
all the measurements routinely used in physics, including
measures of mass, temperature, and various aspects of
electricity and magnetism, rest ultimately on measures of
space (lengths) and time (duration). Sensations cannot be
measured, he argued, because there are no agreed-upon
units: I cannot say that a sensation of strength "4" is
compared with 4 units of strength " 1 , " because the latter
is impossible to define, and moreover I cannot show
ostensively how a sensation of strength "4" can be com-
posed of 4 units of strength " 1 . " This is to be contrasted
with the case for length, where I can prove that a line is 4
metres long by laying 4 one-metre-long sticks alongside it
in a chain. For von Kries, a number assigned to a sensa-
tion magnitude had a status different from that of a
number assigned to length or time: Whereas Fechner had
considered direct magnitude estimates as poor measures
of sensation strength, von Kries argued that sensation
strength could not be measured at all, either by sum-
mated jnds or by magnitude estimates. Any number
assigned to a sensation strength, for von Kries, was no
more informative than a word like "bright" or "dazzling"
or "glaring." True, these words indicated physiological
byproducts associated with particular stimulus inten-
sities, but number ratings were believed by von Kries to
have little extra value compared with such words, and
should certainly not be used as a basis for a "scale" of
sensation strength that could have the same scientific
reliability as a scale of length.
Fechner replied to this argument in his Revision der
Hauptpuncte der Psychophysik (1882, pp. 321-24). He
clarified a point about von Kries's claim: The reason that
length and time can be "measured" so easily is that we are
using measuring devices of the same kind as the magni-
tude to be measured, lengths to measure lengths, seconds
to measure seconds, and so on. Von Kries objected that
psychophysics offered no similar advantage - sensory
magnitudes cannot be measured in terms of sensory
magnitudes. Fechner replied that this claim runs counter
to the evidence: Astronomers can measure star-
brightnesses, by comparing them with other star-
brightnesses, and Delboeuf's subjects could indicate a
greyness intermediate between other greynesses. Fech-
ner therefore felt that von Kries's argument could be
dismissed, but it is clear that he considered it an attack on
the very heart of psychophysics. If von Kries was right, he
maintained, then we would have to abandon psycho-
physics, with the serious result that we would be back in a
wasteland where psychology had no scientific status and
no body of measurements to contemplate. We should

remember that Fechner wrote this before Ebbinghaus,
Pavlov, or Freud had made their discoveries.
It is perhaps appropriate here to mention that Stevens
(1957) also attacked Fechner's principles of sensory mea-
surement, but on grounds different from those of von
Kries. For Stevens, it was Fechner's choice of the "just
noticeable difference" as a unit of sensation strength that
could not be tolerated. The just noticeable difference is an
index of the smallest sensation increment that can be
detected, but no good system of measurement, according
to Stevens, can be based on a unit that simply reflects the
"resolving power" of the measuring instrument. The units
of length in physics are not dictated by the smallest length
that can be seen in a microscope, and only a bizarre
physics would result if we did choose such a unit as a basis
for length measurement. The fact that the subjective size
of the jnd, in Stevens's opinion, did not remain constant is
a further argument against the use of the jnd as a unit of
sensation strength. Stevens's argument against Fechner's
choice of unit should be strictly distinguished, however,
from von Kries's argument that sensations cannot be
measured at all in the sense that length and time can be
measured.
The claim that sensations cannot be measured when
"measure" is defined in terms that have been found
suitable for the physical sciences has led to the develop-
ment of a branch of science called "measurement theory."
In the nineteenth century reactions to von Kries's argu-
ment ranged from the supportive (it was not only von
Kries who objected to the concept of "sensation inten-
sity") to the far-sighted, such as the views of Ernst Mach,
who saw that the problem raised by Fechner had implica-
tions not only for psychology but for measurement theory
in general. It was Mach who claimed that measurements
themselves fundamentally depended upon sensations:
The measurement of length or time is only possible
because the physicist can see that a distance of three
metres is three times as long as a distance of one metre;
the measurement of temperature is chosen to be based on
length (the length of a column of mercury) because that is
a sensation conveniently related to other sensations. The
measurement of temperature could have been based
upon time (e.g., the time to boil a certain volume of water)
but length was found more convenient for practical pur-
poses. A measure of a magnitude, then, is a description of
a relationship between the magnitude and a measuring
device; a measure of length is a description of a relation-
ship between a length and a ruler.
Following the same argument, the measurement of a
sensation might be a statement of a relationship between
a sensation and an indicator of sensation, which might be a
stimulus intensity corresponding to that sensation. How-
ever, Mach actually wrote, "there can be no question of an
actual measurement of the sensations; all that can be done
is to characterize them exactly and make an inventory of
them by numerical means" (Mach 1886/1914, p. 81). On
the other hand, Mach's recognition that measurement is
about relationships between sense-data raised the ques-
tion of whether measurements necessarily concerned
"objective" relationships in the spatiotemporal domain.
The story of how Fechner's views gave rise to consider-
able controversy in the nineteenth century, culminating
in the expression of Mach's opinions, which in turn
influenced Einstein, is told by Heidelberger (1993a).
Murray: History of psychophysics
In the twentieth century it was of course Stevens (1951)
who introduced the distinction between ordinal, interval,
and ratio scales. In the 1980s Laming (1984; see also
Laming 1991) argued that the difficulty we have in ascrib-
ing absolute category values to individual stimuli suggests
that we do not have an internal representation of those
stimuli which is of such a form that category values can be
read off the internal scale: However, Laming does show
that if on trial n we have a stimulus of a given intensity,
then on trial n + 1 we can judge whether the stimulus is
equal to, less than, or greater than the stimulus on trial n,
with the predictability of the resulting assignations of a
category to the stimulus on trial n + 1 increasing system-
atically with the size of the difference between the stim-
ulus used on trial n and the stimulus used on trial n + 1. In
my opinion, Laming's argument also supports the notion
that limitations of memory determine the ascription of
absolute category values and that scales linking stimulus
intensities with estimates of sensation magnitude may be
subject to variability resulting from memory limitations.
Luce (1972) has proposed that the observer himself is a
measuring device measuring the effects of external stim-
ulus magnitudes upon the timing of impulses in the
nervous system, a view that has an affinity with Mach's
theory of measurement. Luce and Narens (1987) have
developed a new view of measurement theory according
to which items on a continuum can be related to items on a
derived continuum by a relationship that can be de-
scribed as having a certain degree of "homogeneity"; Luce
and Narens show that the measurement of sensation has
properties characteristic of certain degrees of homoge-
neity. In their concluding words: "Processes that allow
behavioral attributes to have strong forms of measure-
ment have been developed and measurements of such
attributes, if they exist, will act in much the same way as
physical units. Moreover, it is mathematically feasible for
them to be combined among themselves and with physi-
cal units in just the same way as physical units combine"
(p. 1552). This contemporary optimism is in stark contrast
to the conclusions of the committee chaired by Ferguson
(1940), who were unable to come to an agreement on
whether sensations were measurable, with some mem-
bers upholding the pessimism of the von Kries tradition.
Falmagne (1985; 1986) has made an effort to show how the
psychometric function, the psychophysical law, signal
detection theory, and conjoint measurement theory can
all be related to fundamental propositions in measure-
ment theory.
It is clear, then, that many issues of interest to contem-
porary ps_ychophysicists were first raised and discussed
with considerable acumen by their nineteenth-century
predecessors. The problems of the psychophysical law,
the general problem of the measurement of sensation,
and the problems associated with both neural and experi-
mental noise were all first approached more than 100
years ago. During the intervening years some questions
persisted in an unresolved state and we now turn to an
examination of these,
2. Contemporary solutions to nineteenth-century
problems
The plan shown in Figure 1 is based on a taxonomy of the
key questions of psychophysics. Other issues were raised
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 127
Murray: History of psychophysics
in the nineteenth century which require further historical
discussion; we will now consider two such issues. The first
concerns the nature of the methodology of psychophysics:
Granted that we must measure thresholds or "sensation
strengths," what are the methods whereby we do so? The
second concerns the question of whether a decision about
the nature of the psychophysical law can be based on
information about whether we evaluate differences as
opposed to ratios of sensation strengths of spatially adja-
cent stimuli. Fechner was probably the first to raise each
issue. He himself used three psychophysical methods for
determining thresholds (the method of right and wrong
cases, the method of average error, the method of the just
noticeable difference) and, at the end of his life, he
strongly urged the adoption of Plateau's equisection
method. Fechner also laid considerable stress on the view
that according to a logarithmic law subjects judge differ-
ences between sensations rather than ratios.
2.1. The psychophysical methods
Fechner's three methods were spelled out in The ele-
ments ofpsychophysics; in his later writings he elaborated
considerably on experimental problems associated with
each method. In particular, in hisfinalbook, the Revision
of 1882, he wrote over 100 pages on his three methods,
including a discussion of the treatment of constant errors.
At the same time, Plateau's method of equisection was
sponsored by Fechner and, as mentioned earlier, all
these methods were discussed in detail in Titchener's
scholarly Experimental psychology (1901/1905). Subse-
quent manuals presented the methods in cookbook form
for students (many students in thefirsttwo decades of this
century learned about normal distributions, not from
statistics courses, but from classes on fitting the distribu-
tion to the psychometric function). By the time of Wood-
worth and Schlosberg's Experimental psychology of 1954,
Fechner's three methods had acquired new names. The
method of right and wrong cases became the "method of
frequency" or "the method of constant stimuli." The
method of average error became the "method of adjust-
ment." The method of the "just noticeable difference"
became the "method of limits." Added to these were the
magnitude estimation, ratio estimation, and magnitude
production methods of Stevens and the paired compari-
sons method of Thurstone, methods designed for the
calculation of scales of sensation strength rather than for
the determination of thresholds.
These various methods can be classified in a three-
dimensional matrix with each falling into a particular cell.
Tables 3 and 4 show the matrix. Table 3 involves judg-
ments that are nonverbal (e.g., setting a comparison tone
so that it sounds just noticeably different from a standard
tone, an example of the method of adjustment), whereas
Table 4 involves judgments that are verbal (e.g., giving a
numerical magnitude to the intensity of a tone). Apart
from the dimension of nonverbal/verbal responses, the
other two dimensions are (a) the number of stimuli in-
volved in the comparison task, shown as rows in Tables 3
and 4, and (b) the permissible range of responses in the
reply, ranging from two possible responses ("is A bigger
than B, yes or no?") to very many (as in the method of
adjustment, where the range of possible responses is
determined by the sensitivity of the apparatus). The
128 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
range of responses is depicted by the columns in Tables 3
and 4. The amount of feedback given to the subjects may
vary between methods, as will, of course, the degree of
subjective difficulty of the task; it is unlikely that either of
these variables is mapped in a one-to-one fashion on the
rows (number of stimuli) or columns (range of possible
responses) of this scheme. Most absolute threshold mea-
surement tasks involve two stimuli (stimulus A on a
background G) whereas most differential threshold tasks
involve three (stimulus A and nearly identical stimulus B
on a common background G). Measures involving only
two responses tend to be used for the determination of
thresholds and methods involving a wide range of re-
sponses tend to be used for the construction of scales, but
there are exceptions to this rule: A scale may be based on
the determination of jnds and thresholds may be estab-
lished by the method of adjustment. There are intermedi-
ate cases such as "yes'V'no" judgment tasks with a third
permissible response of "doubtful."
The rows represent the number of stimuli involved in
psychological tasks and it emerges that, with the excep-
tion of tasks involving the rarely used Ganzfeld, all tasks
involve stimuli on backgrounds. Even in absolute thresh-
old tasks, near-total silence is a background, as is near-
total blackness. As we have seen, in the nineteenth
century the distinction was made between background
noise within the subject and background noise related to
the stimulus. It has been conjectured that in near-total
blackness absolute threshold judgments of the presence
of a spot of light are probably more influenced by within-
subject neurological noise than when signals are pre-
sented against a noisy physical background (Cornsweet
1970, Ch. 4). More generally, since every stimulus is
judged against a background, one might raise the ques-
tion of whether sensation intensity is a valid description of
whatever empirical responses are obtained: Should we
not, rather, speak of "sensation distinctiveness"? The
hard-of-hearing person has more difficulty hearing a voice
of intensity / in a crowded room than he has hearing a
voice of the same intensity / in a quiet room; I know one
hard-of-hearing person who dislikes watching films from
the 1930s because he has difficulty hearing what is said
against the noisy sound track. What is consciously per-
ceived seems to depend not on signal intensity alone but
on the signal-to-noise ratio. Whether "sensation dis-
tinctiveness" should replace "sensation intensity" as a
general term might depend on the level of operation of
the brain/consciousness continuum being considered. At
an initial neurelectric level (the receptor level) the inten-
sity of the stimulus is probably coded separately from that
of its background but as the information traverses the
system the signal-to-background ratio may become the
datum that is ultimately reflected in conscious experi-
ence. The distinctiveness of a sensation certainly interacts
with other processes determining conscious experience.
For example, research on human memory has indicated
that if a list of seven visually presented letters such as
BDF . . . is spoken aloud as "bee, dee, eff . . .," imme-
diate recall is more than twice as accurate as when the
letters are pronounced as consonants plus schwas only,
"buh, duh, fuh . . . " (Murray 1967). In the latter case,
distinctive vowel cues are lost, not to mention distinctive-
ness related to long-term experience with "bee, dee,
eff. . . . " The question of the relationship between per-
 Murray: History of psychophysics

Table 3. A classification of psychophysical methods using nonverbal responses

Range of possible responses in psychophysical task (nonverbal)

No of stimuli
in display to Continuous range
be judged ->(H(SH 2-choice nchoice (n >2) of choices

One Background G
Two A on G 1. Adjust C on G
(A barely discrim- to match A on G
inable) 2. Adjust A to be
just noticeable
A on G Pick out C from 2 Pick out C from n 1. Adjust C on G
(A clearly discrim- alternatives that alternatives that to match A on G
inable) match A match A ("magnitude pro-
duction")
2. Adjust C on G
to be a multiple
or fraction of A
("ratio produc-
tion")
Three A and B on G Adjust A to be
(A on B; A barely just noticeably dif-
discriminable from ferent from B
B)
(A near B; A
barely discrimin-
able from B)
A and B on G 1. Select C on G
(A, B clearly dis- that is "in be-
criminable) tween"
A and B on G
("bisection")
2. Select C and D
on G to match A
and B on G either
directly, or in
terms of an equal
difference, or in
terms of an equal
ratto
Four or more A, B . . . on G Arrange stimuli
(A, B . . . clearly on a scale of
discriminable) equal-appearing
sense distances

Note: Modern psychophysical methods arranged (a) according to how many stimuli there are in the display
(rows) and (b) according to the range of possible responses that can be given (columns). Nonverbal responses
consist of either pointing to a stimulus or adjusting a stimulus directly. Empty cells reflect cases where it is
impracticable to obtain responses of the type desired (e.g., "adjust a stimulus to look like this Ganzfeld") or
where no data have been found using that particular combination of stimuli and responses.

ceived distinctiveness (subjective contrast) and the physi-
cal difference between target stimulus and background
will be discussed in the next section. We should also be
reserved in any welcome for the term "sensation dis-
tinctiveness" as opposed to "sensation intensity" because
Schmidt (1991) has shown that loose applications of the
word "distinctiveness" have generated many problems in
the theoretical interpretation of memory data.
Another question raised but not answered by this
classification scheme is the following: We can take re-
sponses generated to the stimulus arrays shown in the
rows of Tables 3 and 4 and this constitutes a body of factual
information. This same body of information might then be
related on rational grounds to the internal scalable magni-
tudes of sensation strength or distinctiveness. Engineers
also make use of the concept of subjective sensation
strength (Indow 1989). Whether such an expanded body
of information is of more value in a scientific psychology

BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 129

Murray: History of psychophysics

Table 4. Modern psychophysical methods arranged as in Table 3 but for verbal responses

Range of possible responses in psychophysical task (nonverbal)

No of stimuli in
display to be Continuous range
judged 2-choice nchoice (n > 2) of choices

One Background G Category judg- "Magnitude esti-

ment of Ganzfeld mate" of Ganzfeld
Two A on G Is A detectable? Canonical ratings
(A barely discrim- ("right or wrong of "sensation" of A
inable) cases")
A on G Is A "X" or "Y"? Category judg- "Magnitude
(A clearly discrim- ment of A estimate"
inable) (with or without
modulus)
Three A and B on G Increment detec- As for 2-choice,
(A on B; A barely tion: "Yes" or "no" but allowing
discriminable (right or wrong "doubtful" judg-
from B) cases) ments
(A near B; A Difference detec-
barely discrimin- tion: "Yes" or "no"
able from B (right or wrong
cases)
A and B on G Is A > B? Category judg- Direct judgment
(A, B clearly dis- ("Paired-com- ment of ratio of A of ratio or dif-
criminable) parisons") to B or difference ference
between A and B
Four or more A and B on G Does (A-B)
C and D on G appear > (C-D)?
Does (AIB)
appear > (CID)?
A, B . . . on G Magnitude esti-
(A, B . . . clearly mates of each of
discriminable) A, B . . .

Note: Verbal responses can range from yes/no judgments, through category judgments where perhaps seven
possible answers are available, to direct estimates where there is no restriction on possible responses other
than the stipulation they may be ordered. The special case where yes/no judgments are also assigned confi-
dence ratings is called "canonical ratings" following Sakitt (1974).
() = neurelectric function; (S) = subjective sensation.

than is the original body of information has been a contro-
versial question. For example, Savage (1970) claimed that
the expanded body of information lacked scientific value
because it depended on introspective judgment. My
feeling is that if we want a science in which assertions
about individual psychological experiences (e.g., "I don't
like stimulus X . . .") are to be meaningfully related to
others (". . . because stimulus X feels too intense"), it is
difficult to see how we could do without the expanded
body of information. Roskam (1991) expresses a similar
view. But in a behaviorist world where subjective experi-
ence is considered unimportant, the body of information
produced from experiments such as those classified in
Tables 3 and 4 would suffice. I think most psychophysi-
cists do want to talk about psychological experience and
that therefore responses in psychophysical tasks will con-
tinue to be related to conjectures about sensation
strength or distinctiveness in the foreseeable future.
Moreover, the question of how measures of sensation
strength are related to stimulus variables and other cogni-
tive variables will continue to be a major one in psycho-
physics.
In the past several years, attempts have been made to
tease apart the effects of the physical stimulus / on the
subjective sensation S from the effects of the subjective
sensation on some other overt response. A stimulus of a
given physical intensity gives rise to a neurelectric func-
tion E [let E =_/(/)], which in turn gives rise to a sensation
[let S = g(E)], which in turn gives rise to a nonverbal or
verbal response R [let R = h(S)]. That is, R = h(g(f{I))). As
Shepard (1981) put it, there is no way whereby, from
knowing an equation connecting R and /, we can sep-
arately estimate / , g, and h, although through experi-
ments we may try to evaluate these functions in isolation.

130 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1


Treisman (1964a) was one of the first to try to isolate the
influence of stimulus magnitude on the neurelectric re-
sponse, contrasted with the influence of the latter on the
subject's overt response. Rule et al. (1974) offered a two-
stage analysis of magnitude estimates with both an input
function (relating stimulus brightness or darkness to an
internal scale) and an output function (relating the inter-
nal scale to magnitude estimates). Marks (1979) offered an
analysis of loudness judgments in terms of separate stages
of processing. Wasserman and Wang-Bennett (1989)
showed that one can measure neurelectric functions,
investigating E = J[I), and that one can also collect
magnitude estimates investigating R = h{g(EJ). They
found that there was a good correlation between E and R
for dark-adapted eyes, but a poor correlation for light-
adapted eyes.
Another question on which there has been controversy
concerns the "anchoring" of external stimuli to sensation
strength by a psychophysical law. Warren (1981) has
argued that our judgments about stimuli are based on
evaluations founded on experience with the stimulus and
its context. For example, we may judge loudness as if it
were a reflection of the distance of the sound source from
the subject, or brightness as if it were an index of the
distance of the visual target from the light source. Contro-
versy also exists concerning the anchoring of the sensation
strength to the response: Poulton (1968) disagreed that
sensation strength was anchored to response output by a
simple linear rule and claimed that there are a large
number of sources of response bias that determine how
we make such judgments as magnitude estimates. The
writings of authors such as Warren (1981) and Poulton
(1968) can be fitted readily into the present perspective if
we think of them as being skeptical about the need for
sensation strength as a tertium quid in the equation
relating stimulus intensity to overt response. Moreover,
N. H. Anderson (1970) has claimed that there are three
quite separate issues: the scaling of what we have called /,
the scaling of S, and the nature of the psychophysical law.
He believes that a concerted attack on all three may be
made by way of his "information integration theory," in
which subjects judge relationships (such as averages)
between two or more stimuli directly. Marks (1974) has
argued that a scale of sensory magnitude should not be
confused with a scale of sensory dissimilarity; the former
may best reflect ratio judgments and the latter may best
reflect interval judgments, both following a psychophysi-
cal power law.
More generally, the study of perception requires that a
distinction be made between the subjective aspect or pole
of a sensory experience and the objective pole, where the
sensory experience is attributed to an event in the physi-
cal world. Scaling of visual and auditory sensations, ac-
cording to Krueger (1989r) is possible without reference
to the objective pole, a statement that Warren might
contest. Nevertheless, the senses differ in their perceived
"subjectiveness," from pain (which is strongly subjective)
to vision and audition (where it is difficult to separate
subjective from objective). There is a need for a greater
specification of the goal of psychophysical scaling in the
light of these differences. The distinction between the
subjective and objective poles of experience was made
particularly by Katz, in his book The world of touch
(1925/1989, translated by Krueger).
Murray: History of psychophysics
Unless the stimuli shown in Tables 3 and 4 were being
used for memory experiments, it may be assumed that
they are being presented simultaneously. Each stimulus
could vary in the number of values it takes on one or more
sensory dimensions. As we move down the rows of Tables
3 and 4 therefore, the stimuli would usually become more
variable in the states they manifest (an exception might be
a comparison between a single target stimulus A on G that
could take, say, 20 possible intensities, as contrasted with
three target stimuli A, B,C on G that could each take only
2 intensities). As a general rule, however, stimulus uncer-
tainty would increase as we move down the rows; and as
we move left to right across the columns, response uncer-
tainty would certainly increase. The more the stimulus
and response uncertainty could vary, the more sensitive
would be a measure of information transmitted, as calcu-
lated by Garner (1962). Since each cell contains a "psycho-
physical method," these methods could be ordered in
terms of their sensitivity to measures of information
transmission. To make the idea more concrete: A simple
"yes" response to the question "Is A bigger than B?"
would tell us little about the size of the subjective differ-
ence between A and B; a ratio estimate "How many times
is A greater than B?" would tell us more.
2.2. How much evidence Is there that "differences"
or "ratios" between sensations do not change
as overall intensity changes?
We have seen that in the nineteenth century this question
was taken to be a key one insofar as it allowed a discrimi-
nation to be made between three of the main contenders
for a psychological law: According to a logarithmic law,
differences between sensations should look the same if
overall illumination (for instance) changes. A power law
suggests that ratios between sensations will be preserved
if overall illumination increases. And Delboeuf's Law
involving neural noise predicts that neither differences
nor ratios will be preserved. Yet neither Krueger (1989a)
nor any of his commentators (except the present) author
Murray (1989) mentioned this question. Is it still a valid
issue? We noted Birnbaum's evidence that perhaps sub-
jects always judge differences even when asked to judge
ratios, and that differences are easier to judge than ratios
on some sensory dimensions, but there is also other
evidence that needs to be evaluated.
In Fechner's own time, the most important study of the
validity of the method of equisection was that of Lehmann
(1886). Delboeuf's display, as shown in Figure 2(a), had
been derived from the measurement of greynesses that
appeared midway between two other greynesses.
Lehmann argued that Delboeuf's results could not have
been reliable, for a number of reasons, including the fact
that (a) such a display was found to include apparent
contrast-induced circles of greyness (i.e., Mach bands) at
the borders of the concentric circles - instead of seeing
three rings, one saw seven or eight; (b) the display was
almost always viewed against the same background,
which induced different contrast effects upon the three
rings in question; and (c) the middle ring differed in size
from the outer and inner rings, again inducing different
degrees of contrast in each of the latter two. Lehmann
instead arranged three discs adjacent to each other, as
shown in Figure 2(b); he found that the establishment of a
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 131
Murray: History of psychophysics
greyness intermediate between two others depended on
whether the variable grey was beside the other two or
sandwiched in between them; he suggested that this was
because the choice of intermediate greyness was influ-
enced by contrast effects which depend on the surround
of the intermediate greyness. These observations led
Lehmann to devise a new type of display, shown in Figure
2(c): A disc of greyness i is viewed against a background
also of greyness i, while a second disc is viewed against a
background of greyness /. The second disc is adjusted
until its greyness matches that of i; call this adjusted
greyness r; then a measure of the contrast induced by/ is
given by (t r)lr. Lehmann's pioneering experiments on
measuring contrast deserve to be better known. His
conclusion was that Delboeuf's evaluations of intermedi-
ate greynesses were confounded by contrast effects and
hence could not be used to test Delboeuf's theories. The
implication would be that the data shown in our Table 1
are also biassed by contrast effects, but Lehmann never
did the experiment of setting up a given contrast and then
seeing whether the contrast looked changed if the overall
illumination was changed. This experiment, it will be
recalled, had been carried out by Delboeuf. The effects of
Lehmann's criticisms were apparently quite damaging,
however, and the twentieth-century neglect of Del-
boeuf's work may have been a consequence. It should
also be mentioned that Lehmann was not the only
nineteenth-century visual scientist to carry out paramet-
ric studies of contrast: Hess and Pretori (1894) also inves-
tigated brightness contrast and their results have been
shown by Hurvich and Jameson (1966) to be consistent
with some findings about contrast that we will now de-
scribe. (They are also consistent with a mathematical
model of contrast that they offer.)
In the twentieth century a large amount of research has
revealed that the assertions one can make about how a
stimulus X "looks' against a background Y have to be
tempered by distinguishing the effects of the absolute
luminances of lightnesses of X and Y from the effects of the
relationship between X and Y. The relationship between
X and Y can be expressed in terms of the ratio of the
luminances or lightnesses of X and Y and this is called the
"contrast" between X and Y; how it looks may be called
the "subjective contrast" or "distinctiveness.'
If we have only two stimuli, a small-sized X on a large-
sized background Y, and we increase the intensity of Y,
the intensity of X that is required to make it detectable
must increase. According to Watson (1986), there are
three steps that the intensity threshold of X goes through
as the intensity of Y increases; a "linear" phase in which
the threshold intensity of X is uninfluenced by the inten-
sity of Y, a second phase in which the intensity threshold
of X increases as the square root of background intensity
(an instance of the de Vries-Rose Law), and finally a third
phase in which the intensity of X is determined by
Weber's Law (here the intensity of X needed to be
detectable is a constant fraction of the intensity of Y). We
can also express these same results in terms of "contrast
sensitivity": If Weber's Law is being followed, the ratios
of objective luminances of X against Y should be constant
and over the range of Y intensities for which Weber's Law
holds, contrast sensitivity should be constant. (In the
earlier two regimes, contrast sensitivity increases.) Wat-
son (1986, pp. 6-36) presents two graphs based on data
132 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
obtained by deLange (1958), one showing how the inten-
sity threshold of X increases as a function of the intensity
of the background Y, the other showing how contrast
sensitivity for X at first increases then flattens out when
the intensity of Y increases to the level at which Weber's
Law holds. Note that this finding should be considered
independently of the claim that the subjective size of ajnd
is not constant as a function of stimulus intensity.
If the array is now extended to include three stimuli, a
test field T presented against an inducing field I, both
presented against a general background G, we can ex-
plore subjective contrast in the manner pioneered by
Lehmann and by Hess and Pretori. Among the best-
known studies of this kind are those of Diamond (1953)
and Heinemann (1955). An extra stimulus known as the
comparison field is introduced: In Heinemann's study,
the test field was a small disc surrounded by a large
concentric disc that constituted the inducingfield.To the
right of this configuration was an isolated comparison field
against the same dark background as the test and inducing
fields. If test-field luminance is held constant and the
comparison field is adjusted to resemble it as closely as
possible while the luminance of the inducing field is
increased, one finds, from analysing the data obtained
concerning the comparison field, how the test field ap-
pears. If the inducing field is set at zero (i.e., melts into
the dark background), the test-field luminance and the
comparison-field luminance are set equal. As the
inducing-field luminance increases, there is a slight ap-
parent lightening of the test field. But when the lumi-
nance of the inducing field is about half the luminance of
the test field, the test field starts to look darker, and from
then on, as inducing-field luminance increases, test-field
luminance looks darker and darker, indicating increased
subjective contrast.
Heinemann varied the experiment in such a way as to
measure the subjective contrast directly. He held the
comparisonfieldconstant and asked subjects to adjust the
test-field luminance until it looked equal to the compari-
son field, at the same time varying the inducing-field
luminance. As the inducing-field luminance increased,
the luminance of the test field that was needed to equal
the luminance of thefixedcomparison field rose slowly at
first but then quickly in such a way that a linear relation-
ship of slope one was found between the log inducing-
field luminance and the log test-field luminance that was
needed to match the comparison field. This means that,
over a particular range of inducing-field luminances, a
testfieldwill look similar to afixed-comparisonfieldif the
ratio of the inducing-field luminance to the test-field
luminance is kept constant. The apparently unchanging
subjective contrast of the test field against the inducing
field that can emerge if the ratio of the two luminances is
kept constant is known as "lightness constancy."
Two predictions follow from this finding. First, if we
have a three-stimulus display, test field against inducing
field, both on a background, and increase the overall
illumination on the whole display by a factor of x, both
test-field luminance and inducing-field luminance should
be increased by x and the subjective contrast of test field
against inducing field should not change (lightness con-
stancy should be found). This often does occur, as attested
to in reviews by Brown and Mueller (1965), Beck (1972),
and Wyszecki (1986). Second, we could present a test

field surrounded by an inducing field (array A) against a
fixed background and a similar test field surrounded by an
inducing field (array B) adjacent to A on the same back-
ground. We would set B's test-field luminance to be x
times A's test-field luminance, and B's inducing-field
luminance to be x times A's inducing-field luminance. We
could then predict that since the ratios of the two lumi-
nances in each of the two arrays are the same, the
subjective contrast of A and B should be identical: In fact,
the two test fields seem the same. This prediction was
confirmed by Wallach (1948; 1963), who also showed,
however, that the two test fields can appear quite differ-
ent if we change the total background of the two, indicat-
ing that some sort of adaptation level determined by the
whole display has to be taken into account in displays
involving more than three stimulus elements.
The impact of these findings is that within certain limits
"equal stimulus ratios" (objective contrast) determine
equal-appearing subjective contrasts. Yet this finding
cannot be said to favor Plateau or Fechner unequivocally
because it is not clear whether subjective contrast is based
on a ratio of sensations (in which case Plateau's theory of a
power psychophysical function would be supported) or on
a difference between sensations (in which case Fechner's
theory of a logarithmic psychophysical function would be
supported). The best that can be said is that a breakdown
of the rule that equal stimulus ratios yield equivalent
levels of subjective contrast would argue against an accep-
tance of either theory as covering all stimulus ratios.
There is indeed an indication that if the stimuli consist of
greys viewed against a white background and these total
displays are illuminated, lightness constancy is main-
tained for only a fixed range of medium greys. Darker
greys tend to look even darker as overall illumination
increases whereas lighter greys tend to look even lighter
(Jameson & Hurvich 1961; Stevens 1961). Brown and
Mueller (1965) argued that these findings could be pre-
dicted from the contrast data obtained by Heinemann
(1955): if the overall illumination increased a test field
intensity / to xl and an inducing field F to xF, whether or
not there would be a change in apparent contrast between
xl and xF as opposed to / and F would depend on the
absolute values of 7, F, xl and xF. Stevens and Stevens
(1960) were able to offer a single equation relating the
subjective brightness of a given grey (7) to both the
luminance of the white surround (F) and the difference
between the reflectance of the given grey (7) and that of
the white surround; the equation predicted that under
increasing illuminance very dark greys should look
blacker.
Each of the above studies used data based on matching
test fields with comparison fields. The study by Jameson
and Hurvich indicated that a psychophysical generaliza-
tion (such as that equal ratios of objective contrast in an
array correspond to equal ratios of subjective contrast)
might only apply to a fixed range of arrays, such as those
involving medium greys viewed against a white back-
ground.
Converging evidence supporting this view comes from
a series of studies of magnitude estimates of the lightness
of grey papers. Warren and Poulton (1960) presented
subjects with a large grey rectangle viewed alongside a
large white rectangle, both against the background of a
brown table or a sheet of black velvet. The greys had
Murray: History of psychophysics
reflectances of 53%, 27.4%, and 9.43%. Brightness judg-
ments of these greys relative to a value of 100% for "white"
could be predicted by taking 10 times the square root of
the reflectance. Warren (1981) later showed, however,
that this square root law broke down for large, very dark
greys (which were judged as being "blacker" than the
square root law predicted) or large, very light greys.
These very light greys were also judged as being darker
than the square root law predicted, a finding that needs to
be integrated with the finding of Jameson and Hurvich
(1961) that very light greys were judged as "lighter"
(rather than "darker") when the overall illumination on
them increased. As Warren and Poulton (1966) them-
selves discovered, the area of the grey patch is important.
The square root law broke down for very small medium
grey patches viewed against white. Warren (1981) used
large areas of very light grey, whereas Jameson and
Hurvich (1961) used small areas of very light grey, but
there is some general agreement between the contrast
studies and the subjective magnitude studies that the
"distinctiveness" of a grey viewed against a background of
white might be described by a law relating distinctiveness
to objective contrast only for a range of medium greys and
breaks down for very light or very dark greys.
If very light greys and very dark greys do not obey laws
of subjective contrast or subjective magnitude that apply
to medium greys, it would probably be wise to use an
equisection method for the construction of a scale of greys
arranged in equal-appearing steps from white to black.
The first such scale may have been obtained by Delboeuf:
In the paper already referred to, he describes how, using
his formula, he calculated what the mixtures of black and
white should be in such a scale. When he constructed a
model showing the scale as a series of concentric rings
ranging in greyness from white (inner) to black (exterior),
one ring looked wrong and Delboeuf found that he had
miscalculated that value. When the error was corrected,
the scale looked right in natural light, but changing the
intensity of the overall illumination caused the outer (very
dark grey) rings to seem too bright or too dim in contrast
with the rest. Similarly, Lloyd Morgan (1900) reported
that after he had empirically constructed a scale of eight
greys on a disc varying in equal-appearing steps from
white to black, the smoothness with which the greys
shaded into each other was disrupted by increasing the
overall illumination: "In a bright light the contrast be-
tween colors and shades of different intensities is height-
ened. A black and white disc which shows good shading
for a medium illumination fails to grade smoothly in a
strong light" (p. 224). He made this the basis for suggest-
ing a new psychophysical function based on calculating an
equation that fitted the empirical data. The equation took
the form
S = UeMl ~ 1) (13)
where L and M are constants.
A scale now commonly used in practice is that derived
for the Munsell system, as renotated by Newhall et al.
(1943). Every color in the Munsell system has a value on
three dimensions, Munsell value (meaning "lightness" or
"greyness"), hue, and chroma (a measure of depth of hue
related to what is commonly called "saturation"; see
Wyszecki 1986, pp. 9-4 to 9-5). The lightness dimension
takes 11 values, 0 (black) to 10 (white). The 11 values are
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 133
Murray: History of psychophysics
supposed to appear equidistant from each other in light-
ness when viewed under normal illumination, but they
are supposed to appear relatively unchanged if the lumi-
nance of the overall illumination is changed. Priest et al.
(1920) derived an equation relating lightness value to
luminance value by a square root function that best fitted
the Munsell values when the latter were viewed against a
white background; the later renotation by Newhall et al.
involved a polynomial equation that fitted the Munsell
values viewed against a medium grey background. Rich-
ter (1953) used a Delboeuf-type equation that fitted the
grey scale of the DIN color chart best when the greys
were viewed against a medium grey background. A re-
view of formulae relating scaled lightness values to lumi-
nance values by various authors will be found in Wyszecki
(1986, pp. 9-40 to 9-44). As noted, such laws usually take
the form of power functions but Foss et al. (1944) found
that a logarithmic law applied well to grey paper chips
viewed against a grey background close to them in
luminance.
The extension of a theory of subjective contrast to cover
multisurface displays is too large a topic to survey here
but we note that Helson's adaptation level theory was
developed in response to the fact that a grey patch viewed
against a white background changed not only qualitatively
but also quantitatively as overall illumination varied: A
grey patch viewed under red light appeared either red-
dish grey, grey, or greenish grey depending on the
luminance of the overall illumination. Gilchrist (1980) has
suggested that individual edges in a multistimuius three-
dimensional array may arise either from boundaries be-
tween two physically different surfaces or from shadows
cast across a given surface, and that the brain integrates
information from these two sources to provide the per-
ception of the three-dimensional world of surfaces and
shadows. Gilchrist has argued that theories of lightness
contrast that depend only on predictions based on esti-
mates of retinal lightness overlook the fact that the subject
also perceives the overall level of illumination.
Over a restricted range of ratios then, it would appear
that equal stimulus ratios yield equal subjective contrasts
ostensibly supporting a power law, but not inconsistent
with a logarithmic law in which contrast is presumed to be
based on differences between sensations. Fechner as-
sumed that a just noticeable difference between two
adjacent stimuli actually did reflect a difference between
sensations. But since one of the stimuli has intensity /,
and the other, according to Weber's Law, has intensity I
+ A/, which equals /(I + k), then an adherent of a power
law could say that if the just noticeable difference still
looks just noticeably different if the overall illumination is
changed, this result is consistent with a power law as well
as with a logarithmic law. A change in overall illumination
would still preserve the ratio 1/1(1 + k). Fechner himself
thought that when a cloud, just noticeably different from
the background, still looks just noticeably different when
viewed through a dark glass, this is evidence for a log-
arithmic law; and it is worth observing that a logarithmic
law was a good predictor of the contrast of Munsell chips
viewed against a background of similar luminance. On the
other hand, Delboeuf found that if he arranged a grey to
be just slightly different from a white background, it did
not always remain distinguishably different if overall
illumination changed.
134 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
In the previous section, it was suggested that stimulus
magnitude might not necessarily be isomorphic with
stimulus distinctiveness. The attempt to describe stim-
ulus distinctiveness in terms of a ratio between the
magnitudes of subjective sensations arising from spatially
adjacent stimulus luminances (Plateau) or in terms of an
arithmetical difference between their subjective magni-
tudes (Fechner) has led to a complicated series of events
whose denouement appears to have been the demonstra-
tion that neither a power law nor a logarithmic law can
predict how all contrasts will look under increasing, or
decreasing, illumination. Yet subjective contrast may not
always be reducible to a ratio of adjacent sensations each
of which has a "strength." Preliminary evidence that
contrast is an irreducible psychological datum comes from
a comparison of intensity increment detection tasks with
contrast increment detection tasks. Intensity increment
detection tasks usually yield results predicted by Weber's
Law except at extreme levels of intensity. But Legge and
Foley (1980) have taken as a reference a stimulus of a
given level of contrast (a 2 cycles/degree sine wave
grating with a mean luminance of 200 cd/m2), with the
task being to distinguish between this and a grating of
slightly greater contrast. As the contrast of the reference
stimulus increased, the contrast of the just noticeably
different stimulus also increased, but it did not follow
Weber's Law: If the contrast is labelled C, AC/C was not
constant but decreased as a function of C. Since the
background field involved gratings with more than two
edges in the field itself, field properties of the stimulus
may have played a part in determining these results.
Twentieth-century work accordingly indicates that
over a range of medium-level contrast situations (e.g.,
medium greys viewed against medium grey inducing
fields) equal degrees of subjective contrast can result from
equal ratios of objective lightness or luminance. This
finding is consistent with the idea that, over a medium
range of brightness or luminances, we perceive ratios
rather than differences between objective lightness or
luminances as a power law predicts; but a logarithmic law
cannot be ruled out if contrast judgments are based on
differences between sensations. Just as Weber's Law
breaks down at extreme values of intensity, however, so
might the power law break down for just noticeable
contrasts or very stark contrasts. When Delboeuf claimed
that his law predicted changes in contrast with changes in
overall illumination, it was at the extremes (dark greys)
that the changes were observed. Moreover, the contrast
between adjacent stimulus elements might be affected by
field properties of the stimulus configuration, as
Lehmann proved and as Legge and Foley'sfindingswith
the contrast of gratings indicate. A more general theory
along the lines of adaptation level theory would be neces-
sary if all the data on contrast are to be assimilated into a
broad theoretical framework.
3. Conclusion
The first half of this target article reviewed the research
that set the scene for two lines of psychophysical enquiry,
the search for a psychophysical law and the measurement
of sensation magnitude by a signal-to-noise relationship.
To the evidence cited, new evidence is continually being

added. The main criticisms of Fechner's Law, with some
newer annotations, were:
(1) Fechner's Law assumes a constant AS for all values
of S. Following Hering, Stevens showed that a power law
could predict a AS that increased as S increased. It should
be noted, however, that Norwich (1987) has developed a
new kind of psychophysical law based on information-
theoretical principles in which it is assumed that AS need
not vary with S, and which still fits the available data.
(2) Fechner's Law is based on the jnd as a unit, which is
a poor kind of unit for a scale. Stevens claimed that the
power law was superior because it could be based on
direct magnitude estimates in which the unit could be set
as a particular sensation strength to which other sensation
strengths could be related (e.g., one "sone" is defined as
the loudness of a 1,000-Hz tone at 40dB above absolute
threshold; Stevens 1936). According to Laming (1989),
however, Stevens's claim that magnitude estimates yield
a power psychophysical law depends on Stevens's having
relied on subjects who had been coached to provide ratio
judgments; moreover, in one case of category judgments
(Montgomery 1975), the data seem to follow a logarithmic
law (Figure 3 of Laming 1989). In Continuing Commen-
tary on Krueger's (1989a) target article, Laming (1991)
argues that Fechner's and Stevens's laws describe differ-
ent types of relationships between I and R, with Stevens's
Law reflecting the operation of certain response biasses.
On the other hand, Zwislocki and Goodman (1980) asked
for magnitude estimates (numbers) or magnitude produc-
tions of tones varying in intensity with respect to specific
sensation levels; there was no bias towards ratio judg-
ments, and they still obtained data well fitted by power
functions.
The power law, which seems to have been applicable to
the prediction of subjective contrast over a wide range of
luminances, does not necessarily apply in the neur-
electric domain, even though Stevens (1970) claimed that
it did. Laming (1989) has argued that direct magnitude
estimates do not correlate with neurelectric measures
taken from human subjects (Knibestol & Vallbo 1980) and
that there is no obvious correlation between power law
exponents obtained from a variety of experiments and the
Weber fractions for the sensory dimensions involved in
those experiments (see Figure 4 in Laming 1989). If the
power law arises from a combination of Weber's Law and
Ekman's Law (as stated in sect. 1.3.4, para. 2), there
should be a correlation between exponents and Weber
fractions. Teghtsoonian (1971) had claimed that the avail-
able data supported such a correlation, but Laming (1989)
suggested that some of Teghtsoonian's data were of ques-
tionable value in this context.
The current fashion involves a variety of escapes from
the Scylla and Charybdis of Fechner and Stevens, includ-
ing (a) attempts to modify Stevens (Krueger 1989a; 1989r);
(b) deriving new psychophysical laws such as those of
Norwich (1987, which involve constants such that for
large values of/ Fechner's Law holds and for small values
of / a power law holds; for intermediate values both laws
fit quite closely) or of Gregson (1991, who "owes nothing
to Stevens or Weber," p. 188); (c) asserting that either
Fechner's Law or a power law can form the basis of a
sensation scale (on such a scale, if sensation M is placed in
between sensations L and IV, the scale will follow Fech-
ner's Law if M is based on the arithmetic mean of L and IV
Murray: History of psychophysics
whereas it will follow the power law if M is based on the
geometric mean of L and N (Falmagne 1985); and (d)
asserting that the simultaneous claims of Fechner and
Stevens to measure sensation are entirely coincidental
(Laming 1989, p. 281; 1991).
On the other hand, the central notion of Fechner's
inner psychophysics, which is that excitation may be
processed in the nervous system to a considerable degree
before it arouses a conscious experience, is in line with a
growing body of experimental evidence. In Fechner's
own time, Baxt in 1871 (see Baxt 1982) argued that a
briefly displayed visual stimulus could be prevented from
entering consciousness if it was followed very quickly by a
flash; we now know more about backward masking and
are ready to assert that intraretinal events, including
temporal luminance summation, are also involved in the
process (Eriksen 1966). Baxt's old idea is currently being
recirculated, however, given the suggestion that word
identification can be facilitated if it is preceded by an
associated word that is masked to the extent that subjects
are not consciously aware that the associated word en-
tered memory (Marcel 1983). Scheerer has also pointed to
evidence from Libet et al. (1967) that "electrical stimula-
tion of the primary somato-sensory cortex [in humans]
may lead to neural responses without any conscious
correlate; as predicted by Fechner, the 'stimulus thresh-
old' and the 'consciousness threshold' are not identical"
(Scheerer 1987, p. 201). [See also Libet: "Unconscious
Cerebral Initiative and the Role of Conscious Will in
Voluntary Action" BBS 8(4) 1985, and Holender: "Seman-
tic Activation Without Conscious Identification in Di-
chotic Listening, Parafoveal Vision, and Visual Masking:
A Survey and Appraisal" BBS 9(1) 1986.]
Whenever an issue is controversial it generates writ-
ings whose volume is inversely related to the degree of
agreement that may exist on the subject. Or, as I put it
elsewhere, "what can be said clearly can be said briefly"
(Murray 1985, p. 237). In this review we have run persis-
tently into two problems which have resisted easy solu-
tion. One is the question of whether sensation can be
measured, along with the related question of whether we
need sensation strength in a more comprehensive psy-
chology. Another is the question of whether, when we
estimate a grey that looks half way between two other
greys, we are basing our judgment on stimulus ratios or
stimulus differences: introspection does not yield an ob-
vious answer. So long as these questions represent slip-
pery ground, psychophysics will continue to be contro-
versial and sensory psychologists will turn with relief to
empirical questions concerning the measurement of
thresholds, contrasts, and so on, because they represent
terra firma. There is an even more treacherous bog
ahead, however, because if it is argued that sensation
strength is an internal scalable magnitude that resists
measurement then other internal scalable magnitudes,
such as habit strength, associative strength, and degrees
of preference, may also lose their privileged status as
kernel concepts in a theoretical psychology that claims to
be based on mathematics. Psychophysics raises problems
for the philosophy of psychological science quite as much
as it raises problems for the sensory psychologist. It is not
surprising that Fechner spent so much time trying to
answer his critics: 100 years later the problems have not
gone away, but we are now in a better position to appreci-
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 135
Murray: History of psychophysics
ate the far-sighted sophistication of many of Fechner's
speculations.
ACKNOWLEDGMENTS
This work was supported by Social Sciences and Humanities
Research Council of Canada Grant 410-89-1040 and a grant from
the Queen's University Advisory Research Council. The author
is particularly grateful to Maureen Freedman for research
assistance; to Peter Dodwell for a critical reading of the first draft
of the paper; and to all the BBS referees, who gave invaluable
suggestions for improvements. Requests for reprints should be
sent to David J. Murray, Department of Psychology, Queens
University, Kingston, Ontario, Canada K7L 3N6.
NOTES
1. Krueger (1991, p. 197) has raised the question whether
amplitude or power should serve as a measure of the intensity of
the physical stimulus. Fechner (1860/1964, vol. 2) wrote of an
intensity as being caused by a body of mass m which in a given
time period moves with a velocity v so that its energy is mo 2 . Let
b be the velocity associated with no sensation (i.e., the threshold
velocity). In our notation, Fechner proved that if S is the
sensation strength and Fechner's Law is correct then
S = 2k log(u/fo) and dS = 2K(dvlv)
where it and K are constants and dS and do are differentials. He
went on to remark that "it is a matter of indifference as to
whether one uses energy or simply speed in these equations,
because in the latter case, one merely doubles the value of the
constants k and K" (vol. 2, p. 30). He also noted that the mass of
the body did not enter into either equation.
2. Luce and Edwards distinguished between a "sensation
jnd" (AS) and a "stimulus jnd" (A/) and defined a Weber function
as a "function relating stimulus to sensation jnd's." If Weber's
Law is true, then since all sensation jnds are equal by definition,
there is a constant P, so that AS/A/ = P/I. Fechner's "mathe-
matical auxiliary principle" allowed him to replace AS/A/ by
dS/dl - the essence of the principle, according to Luce and
Edwards, was to claim that "what is true for differences as small
as jnd's ought also to be true for all smaller differences and so
true in the limit as they approach zero (differentials)" (p. 225).
But Luce and Edwards stressed that for many Weber functions,
for example (A/// 2 ) = k, AS was not constant for all values of/.
The Weber function (A///) = k, however, is a case in which the
values of AS do remain equal.
3. The question may be asked how well Delboeuf's predic-
tions, as contrasted with Fechner's predictions, fit Delboeuf's
data, and whether a power function can be fitted to those data.
Given Delboeuf's stipulation that, on a scale of sensation S,
then Fechner's Law would predict that
Clog/3 - Clogi2 = Clog/2 - Clog/,
if sensation differences are being judged, or, cancelling out C,
that
/ 3 = antilog(21og/2 - log/,) (14)
An indication of how well an equation predicts the obtained data
can be given by plotting predicted values against obtained
values (here, we used Delboeuf's medians); a perfect match will
give a straight line of slope 1, a zero intercept, and r 2 = 1 (r is the
correlation coefficient relating the predicted to the obtained
data). For Delboeuf's equation (Equation 9), the best-fitting
line relating the obtained data to the predicted Z3 data had as its
own equation
obtained data =1.006 (predicted data) + .237 (r2 = .991)
136 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
Using Fechner's equation (Equation 14), the line relating ob-
tained to predicted data / 3 data had as its equation
obtained data = .997 (predicted data) + 2.24 (r2 = .990)
Delboeuf's equation predicted the data slightly better than did
Fechner's equation, but both predictions were remarkably
good.
If sensation ratios are being judged, then we expect S3/Sz =
SJSi, from which it follows that S 3 = SySy. US = Klc (Equation
10 in text), then K/ = (K/|)2/(K/f), which simplifies to
/ 3 = antilog [(2Clog/2 - Clog/,)/c] (15)
Unlike the case for Delboeuf's equation or for Fechner's equa-
tion, C and K in this case have to be estimated from data. Taking
the first entries in Delboeuf's table and setting S = 1 where / =
9, S = 2 where / = 47, and S = 3 where / = 232, the obtained
data can be fitted by the power function
S = .49Z34
The best-fitting line relating the obtained to the predicted / 3
data using Equation 15 was
obtained data = .98 (predicted data) + 2.58 (r2 = .99)
It might be desired to find K and C, so that
K/ = 2K/ - K/f
But this equality only holds if K = 1 and C = 0; otherwise there
will always be a slight difference between the left and right
sides. If a very low exponent is chosen (e.g., C = .01), the
resulting values of S will be indistinguishable from those pre-
dicted by a logarithmic equation (Krueger 1989a, Figure 1). It
should be noted that K is determined by C and a scaling factor,
because if/A is the stimulus intensity corresponding to a sensa-
tion intensity of 1, then K = l^-
It is of interest that the exponent for the power function fitting
Delboeuf's equisection data was .34, which is almost identical
with an exponent of .33 reported by Teghtsoonian (1971) for
magnitude estimates of the brightness of areas subtending an arc
of at least 5. Since Ward (1972) has shown that the exponent for
magnitude estimates is often twice that derived for the same
sensory dimension by categorization methods, it would seem
unwise to think of equisection methods and categorization
methods as somehow equivalent just because both yield parti-
tion scales.
Figure 3 shows that the log / data of Delboeuf can be
transformed by the Naka-Rushton equation to yield S-values
that fall along a curve that has been found useful for fitting log /
data to neurelectric measures of receptor response (Lipetz
1971). However, Delboeuf did not use the extreme values of
lightness that would have better permitted us to evaluate the
goodness of fit of the Naka-Rushton function.
A logarithmic function predicting equal sensation differences,
a power function predicting equal sensation ratios, and a trans-
formation yielding a curve similar to one found for receptor
potentials could all be applied to Delboeuf's data. This fact
indicates that curve-fitting alone is unlikely to decide between
psychophysical functions as applied to real data in certain cases.
4. S. S. Stevens was aware of this early research and briefly
mentioned it in his 1957 Psychological Review article.
5. It is perhaps appropriate to mention here that several
authors have pointed out that a power law can be derived
without reference to the jnd, including Stevens (1957) and
Yilmaz (1967). For example, according to Yilmaz, if we assume
that to physical variables i, and t2 there correspond sensory
variables sl and s 2 respectively and that the percept relating s,
and s 2 is given by
(S2/S]) = ji^h) (^)

I Open Peer Commentary
log I log Io
Figure 3. A demonstration that a curve that has the shape of
the Naka-Rushton function can be generated from Delboeuf's
data. In the original formulation, which was about the way the
receptor potential Vo grew as the logarithmic of the stimulus
intensity (u>) increased, Vo = .5 + .5 tanh (w w0), where w0
was the log stimulus intensity corresponding to the half-
saturation intensity, of Vo. Here we have simply relabelled the
abscissa showing (w w0) as (log / - log / 0 ) and the Vo ordinate
as sensation strength. The abscissa was arranged so that Vo = 1
corresponded to log 360, the highest possible value of lightness
in Delboeuf's experiment, and Vo = 0 corresponded to a
Delboeuf value of 0. Following the instructions of Lipetz (1971),
a rough curve was drawn through a few data points to estimate
the value of / (corresponding to w0); then the rest of the data
points were calculated from this by transforming (log / log / 0 )
according to the above equation. There are two matters of
interest. First, Delboeuf did not explore extremely dark light-
nesses or extremely light lightnesses, so there are few data at the
ends of the curve. Second, all the points above the level
corresponding to V() = .8 are data for subjects' responses; a few
response data lie below this level, but most of the points below
Vo = .8 reflect lightnesses of stimuli preset by Delboeuf.
where / m e a n s "is a function of," then it can be shown that
s = Bf (17)
where B is a constant and x can be specified in terms of/. As
Yilmaz points out, "the derivation of this first-order law makes
no use of the concept of the just noticeable difference (jnd) or
resolving power. Nor is Weber's Law required" (p. 534). Equa-
tion 16 implies that "equal stimulus ratios imply equal sensation
ratios"; later evidence related to this assertion will be consid-
ered in section 2.2. Luce and Edwards (1958) note that since
jnds are simply cumulated it does not necessarily matter for
practical purposes in constructing a scale whether Weber's Law
is actually valid for the sensory continuum being investigated.
6. Corso (1963) has pointed out that for Fechner (and, later,
Stevens), any change in the point at which the threshold is
measured (e.g., the point at which the subject detects the
stimulus 50% of the time) will change the zero point on the
psychological continuum that is supposed to reflect the physical
continuum. Some theoretical accounts of the psychometric
function, such as Falmagne's (1986), make this point a variable
Commentary /Murray: History of psychophysics
in their model. Corso notes Michels and Helson's (1949) sugges-
tion that replacing / by the adaptation level (as previously
measured) can get round the problem.
Commentary submitted by the qualified professional readership of this
journal will be considered for publication in a later issue as Continuing
Commentary on this article. Integrative overviews and syntheses are
especially encouraged.
Nonconscious sensation
and inner psychophysics
Norman H. Anderson
Department of Psychology, University of California, La Jolla, CA
92093-0109
Electronic mail: nanderson@ucsd.edu
1. Nonconscious sensation can be defined and measured with
the functional measurement methodology of integration psy-
chophysics. Integration psychophysics thus involves a way of
thinking rather different from most versions of traditional psy-
chophysics, which almost define sensation as conscious.
Nonconscious sensation is common. What reaches conscious-
ness is often, perhaps nearly always, an integrated resultant of
multiple stimulus determinants. The multiple effects, or sensa-
tions, evoked by these separate stimulus determinants will be
inaccessible to consciousness in many cases. Traditional psycho-
physics, having largely ignored the integration problem, has
remained largely unconscious of nonconscious sensation. Inte-
gration psychophysics can dissect conscious sensation to deter-
mine - measure - its determinants, both conscious and
nonconscious.
One example is the size-weight illusion, in which the effect of
the visual appearance on felt heaviness is beyond the reach of
consciousness. Because the two determinants are integrated by
an addition rule, both can be measured on true linear scales with
functional measurement.
These separate sensations, it may be emphasized, can have
substantial effects on the conscious sensation. They are not
nonconscious because they are marginal, at or below "thresh-
old," but because they operate at a stage of information process-
ing prior to consciousness.
2. Murray errs in his description of the functional measure-
ment methodology of the theory of information integration.
There are three quite separate issues, it is true, but only one is
correctly given. Contrary to what Murray suggests, measure-
ment of the physical stimulus and determination of the psycho-
physical law have only secondary importance in integration
psychophysics.
Instead, the three issues refer to: (a) sensation prior to
integration, often nonconscious; (b) the integration function
itself, also called the psychological law; and (c) the integrated
resultant, prototypically conscious. All three issues can be
resolved jointly when the psychological law has an algebraic
form. Numerous applications (Anderson 1982; 1990; 1992;
McBride & Anderson 1991) demonstrate the empirical effec-
tiveness of this functional measurement methodology.
The conceptual shift from psychophysical law to psychological
law is essential to integration psychophysics. The psychological
law, being a function of two or more variables, embodies
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 137
Commentary/Murray. History of psychophysics
mathematical constraints that make it possible in principle to
solve the problem of psychological measurement. By a blessing
of Nature, which constructed the mind to follow algebraic rules,
this solution can be realized in practice. Theory and practice are
both illustrated in the cited work.
3. Inner psychophysics is the prime domain of integration
psychophysics. The input variables to the psychological law are
not the physical stimuli but their psychological representations,
that is, their sensation values prior to integration. The integra-
tion function, or psychological law, transforms these separate
sensations into a unitary, conscious response.
This view of inner psychophysics, however, is conceptualized
differently from Fechner's. Fechner did not realize the impor-
tance of integration. This limitation, understandable in his time,
was unfortunately welded into his concept of psychophysical
law, which is a function of a single variable. The conception of
psychophysical law maintained an iron grip on Fechner's suc-
cessors, as Murray's target article so well demonstrates. Hence
the potentialities of integration psychophysics were not appreci-
ated. Numerous workers have studied stimulus integration, it is
true, but this has generally been restricted to limited special
questions and contexts.
4. Murray concludes by emphasizing the importance of mea-
surement for theoretical psychology, noting also that it has
occasioned much controversy. Murray's discouragement, how-
ever, is not warranted. In fact, the problem of psychological
measurement has received a reasonably satisfactory solution in
the functional measurement methodology of integration
psychophysics.
I wish to conclude by invoking Murray's fine aphorism that
what can be said clearly can be said briefly: The conceptual shift
from psychophysical law to psychological law, with the blessing
of Nature, has provided a viable theory of psychological mea-
surement. This theory is not a promissory note; it is solidly
grounded in numerous empirical applications, not only in psy-
chophysics, but also in other domains.
A perspective on psychophysics is not
derived just from the history
of psychophysicists
Gunnar Borg
Division of Perception and Psychophysics, Department of Psychology,
Stockholm University, S-106 91 Stockholm, Sweden
Howard Bartley (1970) pointed out the importance of a good
name for the identification and recognition of a sensory system.
By analogy, we may say that psychophysics has profited by its
name. We must not, however, confuse the name with the
concept. That Fechner described psychophysics as a novel and
precise scientific field and coined the term "psychophysics" does
not entail that the perspective should start with him. Further-
more, it should be broadened to include the theoretically
challenging and practically momentous problems of fundamen-
tal "level determinations."
The main subfields of psychophysics are said to be detection,
identification, discrimination, and scaling; of these areas, scal-
ing is the least exact. In all of these areas foundations can be
traced to well before Fechner. In this respect, Weber's contri-
bution to the then "nameless" subject of psychophysics has been
underrecognized, as has the mathematically unpretentious (but
for human communication so crucial) linguistic categorization
"and rating scaling."
Metaphysical and epistemological ideas about other minds
can be traced back to folk psychology, defined as hypothetical
constructs, and then, using hypotheticodeductive reasoning,
can be specified and tested empirically (Churchland 1990).
Even in the Bible, we find examples of psychophysical thought.
138 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
Take Amos 5:2: "The city that went forth a thousand shall have a
hundred left, and that which went forth a hundred shall have ten
left." Here we have an example of intensity evaluations in
geometric series. According to Daniel Algom (personal commu-
nication), Bernoulli (1738) took this to support his logarithmic
law. Bernoulli should be credited not only for his log law, but
also for his recognition of intensity levels and individual differ-
ences in utility. He (as well as La Place) proposed relative
intensity evaluations founded on individual "maxima." Fechner
(1860) writes about this but does not build further upon this very
important idea. This concept of relative intensity evaluations is
also illustrated in the Bible when Jesus recognizes the subjec-
tive value of the gift by the poor widow.
Many interesting sources for a psychophysical perspective
may be found in the classical literature as well. Take Shake-
speare's A midsummer night's dream, for example. At one point
Lysander cries: "Thy threats have no more strength than her
weak prayers!" And at another Helena comments: "Weigh oath
with oath and you will nothing weigh, your vows to her and me
put in two scales will even weigh, and both as light as tales."
Murray only hints at the problem of fundamental level deter-
minations in one sentence (in sect. 2.1): "because stimulus X
feels too intense." He should have developed this important
problem further. We must search for sources outside the work of
psychophysicists and mathematical psychologists. In everyday
communication and in evaluations of clinical symptoms, we
discuss whether something is strong or weak in an "absolute"
sense.
Both Fechner and Stevens failed to recognize these prob-
lems. It is interesting to see in Stevens's last article, where he
condemns category scaling (Stevens 1971), that when he dis-
cusses individual differences he uses category expressions such
as "very loud" and "slightly painful" to invalidate the results of
his own ratio scaling methods. Furthermore, Stevens errs with
regard to level determinations when he says that the difference
between the two exponents for two observers suggests that the
loudness may be 3,000 times greater for one observer than for
the other. Here he disregards the difficulty with units for
interindividual comparisons.
A possible solution of this "unity problem" is found in the
range model proposed by Borg. According to this model, a
certain extremely strong intensity, such as the strongest per-
ceived exertion, is set subjectively equal (interpersonal sim-
ilarity in sensory perception) and used as a "fixed point." The
intensity of a sensation is then evaluated depending upon its
relative position in each individual's range from "minimum" to
"maximum" (see Borg 1961; 1962; 1990; Marks et al. 1983).
Psychophysics may be looked upon as applied epistemology.
My teacher David Katz (as a student of Husserl) was interested
in problems of interpersonal subjectivity, as was Go'sta Ekman.
Neither of them, however, ever presented a working model for
interpersonal comparisons. Ekman concluded that it was theo-
retically impossible to define a valid unit; two individual curves
could never be drawn in the same diagram (personal communi-
cation; see also Bjorkman & Ekman 1962).
Quine (1960; 1987; 1990) discusses the importance of inter-
subjectivity and "shared stimulus meaning." He points out that
he finds Stevens's finding interesting but that he doesn't know
what to make of it. In the 1987 book, Quine writes: "There is no
place in science for bigness, because of this lack of boundary; but
there is a place for the relation of biggerness. Here we see the
familiar and widely applicable rectification of vagueness: Dis-
claim the vague positive and cleave to the precise comparative."
At the end of section 1.3.3 Murray points out the need for
another psychophysical equation. The first solution to this
problem, to cover both perceptual and physiological functions,
was proposed by Borg (1961; 1962) and Mountcastle (1963), that
is, a power function with two extra constants. Physiological
support for the power law has been provided by the studies on
taste perception by Borg et al. (1967) and on perceived exertion

by Borg (see, e.g., Borg 1978; also Borg et al. 1987). Invariant
relations have been found between heart rate, blood lactate and
ratings (R) of perceived (P) exertion (E) according to the "RPE-
scale" (probably one of the most commonly used psychophysical
methods, with several hundred studies performed in the last
few years and approximately one million people exposed to it
every year), and the "CR-scale' (category scale with ratio prop-
erties, Borg 1982). For interpersonal comparisons a "rating
scale" may be superior to a "ratio scale"!
Guilford (1936) provides a good review of rating scales (with
an example from descriptions of weather in 1805 reported by
Titchener). He did not make the same mistake as Stevens and
most psychophysicists, including Fechner - notable exceptions
are Parducci (1984), Heller and the Wiirzburg group (see, e.g.,
Heller 1990), Borg (1962), and a few others. Guilford under-
stood the importance of "descriptive terms" and "landmarks."
He concludes that the rating methods "have overshadowed the
applications of all the other evaluative methods based upon
personal judgements" (p. 263). But he also points out how
"indispensable" these methods are to theoretical psychology. I
definitely agree with him, but would also like to stress the
importance of using the best properties of the ratio scaling
methods for general, relative intensity evaluations, together
with the best properties of the category-rating methods for level
determinations and interindividual comparisons. This use of
"meaningless" numbers with "meaningful" linguistic terms will
improve the possibilities of understanding human communica-
tion and the problems of intersubjectivity.
The phantom limb extrapolation
Willard L. Brigner
Psychology Department, Appalachian State University, Boone, NC 28608
Electronic mall: brignerwl@conrad.appstate.edu
In tracing the development of Fechner's inner psychophysics,
Murray seems apologetic (oid. sect. 1.2, para. 5). In presenting
signal detection theory as the major development of Fechnerian
inner psychophysics, he notes, no doubt correctly, that there
are only "commonalities of interest" between signal detection
theory and Fechnerian inner psychophysics. In other words,
the Fechnerian tradition should be considered incidental in the
development of signal detection theory and should not be
considered conative. This approach to the topic of inner psycho-
physics seems feckless. Instead, when Murray alludes to the
neurelectric aspects of a stimulus (e.g., sect. 1.2, para. 2) or to
neurelectric measures of receptor response (e.g., Note 3 and
sect. 1.3.4, para. 3), one would hope that the problem at "the
heart of inner psychophysics" (sect. 1.2, para. 2) is to be
addressed, but it never is.
In discussing inner psychophysics, Murray could have cited
some long-standing electrophysiological data which support
Fechner's Law in that the frequency of neural firing has been
shown to be in approximately direct proportion to the log of the
stimulus intensity. This relationship has been demonstrated, for
example, by recording from the nerve fiber of an isolated
oinmatidium in the Limulus eye (RatlifF& Hartline 1959). It has
also been demonstrated by recording from single axons of
stretch receptors of the muscle spindle in the cat (Patton 1960).
Consistent with the power law, in recordings from single nerve
fibers subserving mechanoreceptors in the skin of the monkey,
the log rate of neural firing has been shown to be approximately
directly proportional to the log stimulus pressure (Werner &
Mountcastle 1965). (Note that the question of whether the
appropriate law is logarithmic or a power law extends to electro-
physiological data.)
Of course, the electrophysiological data and the sensation
scaling data offer an interesting implication. The implication is
Commentary I Murray: History of psychophysics
that frequency of neural firing (at the receptor level) and
sensation strength (or psychological magnitude) are inter-
changeable, and an interchangeability of something physiologi-
cal with something psychical suggests inchoate equivalences
regarding, as Murray describes it, "the heart of inner psycho-
physics . . . the 'mind/brain'problem . . ."(sect. 1.2, para. 2).
Of course, Murray's reticence in developing inner psycho-
physics along these lines can be understood, since one rarely
encounters the Cartesian issues of mind-body dualism, innate
ideas, and so on, in the current psychological literature. But
there are exceptions - such as the following.
Based upon work with patients experiencing phantom limbs,
phantom vision, and phantom hearing, Melzack (1992) extrapo-
lates the experience of phantom limbs to the experience of a
phantom body. He writes that "The brain does more than detect
and analyze inputs; it generates perceptual experience . . . [in
fact] the brain generates the experience of the body . . . .
Sensory inputs merely modulate that experience . . . . " (p.
126). Although "the experience of the body" may not be on
Descartes* list of innate ideas, Melzack would seem to be
offering it as a candidate. His brain-generated experience of the
body provides an amusing contrast to Locke's discussion of the
tabula rasa, where there was a body and the phantom was
phenomenal experience. So Murray's development of inner
psychophysics would have been more effective if it had pos-
sessed some of the temerity of the phantom limb extrapolation.'
NOTE
1. A final trivial point: Murray fails to distinguish between border
contrast and Mach bands (oid. sect. 2.2, para. 2). The incre-
ment/decrement in apparent brightness at borders, where there is an
abrupt step in intensity, is usually referred to as border contrast. If the
intensity distribution includes an intensity gradient linking different
levels of intensity, the increments/decrements in apparent brightness at
the gradient end(s) are referred to as Mach bands.
From metaphysics to psychophysics
and statistics
Gerd Gigerenzer
Department of Psychology, University of Chicago, Chicago, IL 60637
Electronic mail: giger@midway.uchicago.edu
David Murray has reminded us that the issues raised by Gustav
Theodor Fechner are still the stock-in-trade of modern psycho-
physical laboratories. I will follow Murray's lead here and ask
the question whether (and how) Fechner's issues and their
modern offspring are related to Fechner's philosophy.
Fechner's metaphysics is a unique blend of monism, panthe-
ism, panpsychism, and indeterminism, for which no single
English word exists. Three books published before the Ele-
mente der Psychophysik (1860), none of which has been trans-
lated into English, contain the elements of Fechner's meta-
physics. In Das Biichlein vom Leben nach dem Tode (first
published in 1836 under the pseudonym Dr. Mises), Fechner
argued that man lives not once, but thrice. The first life is a state
of constant sleep, the second is a periodic alternation between
sleeping and waking, the third is eternal waking. The transition
is through birth and death. To be awake means to be conscious,
and consciousness means that the "so-called psychophysical
activity rises above that degree of strength which is called
threshold" (p. 51). In Nanna (1848) Fechner argued that plants
have souls. In Zend-Avesta (1851) he explained why the earth
and planets have souls as well.
The idea that not just we humans have consciousness flows
from his monism: Mind and body are the same thing, they just
look different as seen from inside and outside, respectively. If
you look into yourself, you perceive psychical processes; if I look
at you, I perceive a material entity. The essential unity of mind
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 139
Commentary/Murray: History of psychophysics
and body implies that not only humans are conscious, but plants
and planets as well. Fechner points to Kepler's similar ideas in
Harmonia mundi, where the earth is a living creature who
alternates between sleeping and waking states. Finally, nature
seen as a whole has a soul as well, which is God. Seen from the
perspective of Fechner's monism, to search for causal laws that
link the psychical to the material or vice versa, is a waste of time.
There are no such laws between mind and body: There is
identity, no succession. Causal laws only hold within each of the
two realms. This brief sketch of Fechner's metaphysics must
suffice for present purposes.
Did aspects of Fechner's scientific work originate in his
metaphysics? If so, does psychophysics still bear these birth-
marks? I will attempt a partial answer by examining Fechner's
two major treatises on psychophysics and statistics.
1. In his Elemente der Psychophysik (1860), Fechner explic-
itly links his psychophysical program to his metaphysical work
(e.g., vol. 2, p. 543). Several writers, from Boring (1929, p. 269)
to Murray (sect. 1.1) have pointed out that Zend-Avesta already
contained Fechner's program of psychophysics. More specifi-
cally, the common interpretation seems to be (a) that Fechner's
program of measuring sensation and determining psychophysi-
cal functions was a means to prove monism, that is, to demon-
strate the essential unity of mind and body through a simple
mathematical equation, and (b) that in Zend-Avesta, Fechner
intuited the logarithmic shape of the psychophysical function
before he did any experiments. I have two comments.
(a) Psychologists may not have embraced Fechner's program
in order to grasp immortality (Fechner's own aim), but they did
nonetheless adopt the program as a means to their own ends.
Does this mean that its scientifically embarrassing origins are
accidental to the history of psychophysics, amusing at best? I
don't think so. Fechner's metaphysics may well have given the
history of psychophysics an impulse in a particular direction.
Here is the argument. There are many possible programs of
psychophysics. Not every program can serve as a means to prove
monism. Programs geared toward identifying psychophysical
functions, whether Fechner's or Stevens's, can serve that end.
Programs that reject a one-to-one relationship between sensa-
tion and stimulus, however, would not have served that purpose
- from Gestalt psychology to Brunswik's (1934) multidimen-
sional psychophysics to the psychology of contextual effects
(e.g., Parducci 1982). That is, if one assumes a parallelism
between brain states and perception, but complex Gestalt laws
rather than a parallelism between these two and the physical
world, then no simple psychophysical function exists. Or, if
contextual effects result from coherence systems and perceptual
compromises, as in Brunswik's multidimensional psycho-
physics, again no single psychophysical function exists (Gig-
erenzer & Murray 1987, Chapter 3). It is, of course, not easy to
say what could count as evidence for the suggested impulse from
metaphysics, and the pointed neglect of contextual effects in
psychophysics was no doubt overdetermined. The best I can do
here is to point to the difficult time contextual effects have had
throughout the history of psychophysics, often ignored or down-
played as mere "nuisances" or "response biases" (on this see
Birnbaum 1989). The questions that define the history of psy-
chophysics seem to exclude contextual effects in favor of psycho-
physical functions.
(b) My second comment also addresses the context of discov-
ery: Where does the logarithmic shape of the psychophysical
function come from? The idea is already in the Zend-Avesta,
although it does not seem necessary for proving monism; a
power law would have done just as well. Murray (1987, p. 77)
has pointed to potential extrametaphysical origins: The log-
arithmic laws formulated by Daniel Bernoulli or by astronomers
may have served Fechner as an analogy. There exists one
further possible origin, which is more closely linked to Fech-
ner's own research. Fechner had spent two years of research
140 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
(1829-1831) on the galvanic battery, which earned him a chair of
physics at the University of Leipzig. Fechner's research was
based on Ohm's path-breaking treatise on that subject. As
Stigler (1986, p. 243) reminds us, Ohm's first paper of 1825
postulated the formula
V = m log (1 + i) (1)
for the relationship between the loss offeree of current V and the
length of wire x. Fechner proposed the same logarithmic rela-
tionship for the psychophysical function,
S = it log fl, (2)
where S is the sensation, R the physical stimulus, and k a
constant value. Thus, the idea of a logarithmic psychophysical
function may well be by analogy to Ohm's law. To summarize
the gist of these two comments: In the history of psychophysics,
Fechner's metaphysics may well have played a role. Meta-
physics has helped to select one class of programs and to exclude
others. Within that selected class, however, an analogy from the
physical sciences seems to be the source of a particular
expression.
2. In his Kollektivmasslehre (posthumously published in
1897), Fechner is silent about any connections with his meta-
physics, except for a hint in the preface (p. vi). In the last
decades of his life, Fechner became reluctant to speak about
metaphysics when addressing a scientific audience, realizing
how unpopular his metaphysical mission was among his fellow
scientists. The Kollektivmasslehre is a theory of frequency
distributions, and its goal is the description oikollectivs (such as
the height of Belgian soldiers or the dimensions of religious
paintings). Its tools are asymmetric generalizations of Gaussian
distributions, measures of central tendencies, variability and
correlation, and tests of statistical independence. In a seminal
article, Heidelberger (1987) has traced its relation to Fechner's
philosophical indeterminism. In Fechner's monism, the mental
is not causally determined by the material, rather, the mental is
indeterminate in several ways, most importantly in the sense of
intrinsic novelty. Novel elements emerge in the mind, leading
to novel consequences. Because mind and matter are just two
appearances of the same entity, true indeterminism - not just
Laplacean ignorance of conditions - also holds in the physical
world. Indeterminism, however, does not preclude the scien-
tific study of nature (a remarkable view to be voiced long before
chance became an indispensable part of physics). The Kollektiv-
masslehre was Fechner's program for studying indeterminate
nature.
Fechner's program is a strictly frequentist theory of proba-
bility, in marked contrast to the classical interpretation of
probability, dominant circa 1660-1840 (Daston 1988). Richard
von Mises, founder of the frequentist theory in our century, is
explicit about the influence Fechner had on him: "his views have
served, at least for me, as a stimulus in developing the new
concept of probability" (von Mises 1928/1957, p. 83). The rest of
the story is well known. Frequency statistics eventually con-
quered psychophysics and all of the experimental social sci-
ences, albeit not in the form of von Mises's large sample
statistics, but through R. A. Fisher's (1935) small sample statis-
tics. By this strange route, metaphysics contributed to the
hardheaded empiricism of logical positivism.
To summarize: Elements of Fechner's scientific work, both in
the Elemente and the Kollektivmasslehre, can be traced back to
Fechner's metaphysics. This source opened new programs, but
may also have inhibited the growth of others. Looking at the
context of discovery, even one as mystical to our eyes as the
present one, can help us understand part of the present as well
as the past of psychophysics, including the origins of the kinds of
question and tools that are still with us.

The head and tail of psychophysical algebra
Robert A. M. Gregson
Department of Psychology, Australian National University, Canberra ACT
2601, Australia
Electronic mall: rag655@cscgpo.anu.edu.au
Murray's scholarly and timely target article does much to put
things back into perspective and reinstates nineteenth-century
contributions later either lost or misattributed. He is to be
warmly thanked.
There are two ways of looking at psychophysics, however;
either as the roots of experimental psychology, which Murray
does meticulously, or as applied mathematics. From the latter
viewpoint, Murray's account is like an artist trying to draw a still
life of a fish from which, unfortunately, his cat has already
chewed off the head and the tail. No matter how accurately he
delineates what remains (which admittedly are the guts of the
fish) essential fish qualities are lost.
The mathematics of Fechner are not sophisticated, even by
the standards of his time. Precursors had used linear differential
equations on psychological topics before him, and he knew of
some. Setting aside some sensation ratio model equations writ-
ten in the mid 1700s from Tartu (Dorpat) in Estonia, the best
starting point is still Herbart's (1812) paper on the strength of a
given impression as a function of its observed duration; this was
obviously well ahead of his contemporaries, both in the detailed
consideration of sensory processes and the use of differential
equations in model construction. The radical step in the argu-
ment, which was apparently dropped from psychophysics for
some time, or rather partialled out to become the separate study
.of reaction times, comes when Herbart states "Both the abso-
lute, and the relative strengths are Functions of Time. The word
'function' is here taken in the mathematical sense, as it means a
changing magnitude which itself depends on another thing
changing." The great interest for the modern mathematical
psychologist is that Herbart used no stochastic theory; the
equations antedate the partition into deterministic and proba-
bilistic parts which Gauss introduced a bit later in his theory of
errors. At the same time, many technical devices used by
Herbart are familiar, and the emphasis on conceptualising
processes as extending in real time, and converging on stability,
is a precursor to aspects of dynamical systems theory that surface
again with Lyapunov's habilitation thesis in 1892 (Goldner
1983).
Contemporary mathematical developments in nonlinear sys-
tems theory (Wiggins 1990) make it possible to reexamine the
fundamental questions which Herbart insightfully addressed;
the computational difficulties that blocked him, and perhaps
dissuaded his successors from exploring the same algebraic
structures, no longer impede or intimidate us.
Obviously Fechner's mathematics were not unique in his own
day. Sprung and Sprung (1987) note that a tradition of "Physik
der inneren Erfahrung" tracing from Euler and Herbart to
Drobisch was alive; indeed Drobisch (who is misspelt as Fro-
bisch in Adler's translation of Fechner's Elemente) had pub-
lished two texts before Fechner. They were Empirische Psycho-
logie in 1842 and Grundlehren der mathematischen Psychologie
in 1850. Fechner was aware of both, and cites Drobisch.
Sometime professor of mathematics at Leipzig, Drobisch (1861)
promptly amended the derivation of Fechner's formulae. Kiilpe
(1893) refers explicitly to this tradition; it is in opposition to the
Kantian stance denying the possibility of psychological mea-
surement which Murray properly associates with von Kries. It is
not even the case that Fechner's extensive treatment of the
Method of Limits is original; the method goes back at least to
1827 with Delezenne, and the early evidence for Weber's Law
rests on data obtained by that method; indeed the two stand or
fall together. Weber used the method in 1831 in Anotationes
Commentary /Murray: History of psychophysics
anatomicae et physiologicae in Wagner's Handworterbuch des
Physiologie, Band III. One must emphasize that Fechner's
mathematics has all its roots in the eighteenth century, and the
contemporary developments in the mid-nineteenth century,
associated with such workers as Clerk Maxwell, were in a
physics which had moved on (Bell 1939). Late twentieth-
century developments - analytically intractable but computable
ones - were obviously not even imagined and yet to come; all
Fechner had was pencil and paper (and presumably log tables).
The surprising and indeed powerfully insightful features of
Herbart's mathematical model are:
(i) The process is in real time.
(ii) The process is bounded, with limits and with internal
inhibition as feedback.
(iii) Both the output of the process and its time course follow
from the same model.
(iv) The concern is with the metric properties of psychologi-
cal variables.
Although Herbart modestly (at the beginning) and less mod-
estly (in the middle, half way through the algebra) makes
reference to his mathematical competence, and although the
exercise ha.s no experimental data to complement it, the formu-
lation is coherent and ties together what later fell apart.
If you look at the literature of contemporary psychophysics
and its applications in cognitive psychology, you will find dis-
tinct and separate treatments of
(i) the psychometric function, with excessively massaged
data plots of stimulus-response mappings,
(ii) response latency frequency distributions, with even more
massaging into skew normal or Pearson type III form, and
(iii) formal treatment of scaling theory, as in the well-known
Foundations of measurement, vols. 1 to 3, 1971-1990; (see
Smith 1991).
We know that Euler published some bizarre mathematical
"proofs" that the soul is not a material substance (Bell 1939) and
this could have created the intellectual licence for Herbart to try
what he did. But Herbart's attempt to solve his algebra with a
series expansion and an approximation to a few terms is Eule-
rian. Today we would do it by simulation and numerical
methods on a computer if there were no ready known analytic
solution, but that was not really practical until the 1960s.
Herbart was 150 years ahead of the hardware he needed. The
important point in the exercise is the date, 1812. (I don't refer to
Napoleon's contemporary retreat from Moscow!) Though Gauss
has reworked de Moivre's 1732 analysis of error in 1809, to get
what we call a normal distribution, it was only in 1827 that Gauss
used the theory of errors (1821, 1823, Theoria combinationis
observationum erroribus minimis obnoxiae) to facilitate the
resurveying of the city of Hannover, having developed it for his
astronomical work. The idea that we might treat the psycho-
physical problem as a linear law with superimposed residual and
normally distributed error is nowhere in Herbart; the 1812
algebra is deterministic but involves something like a closed-
loop recursion.
Now, Gauss coedited a geomagnetism journal with Wilhelm
Weber and collaborated with him up to 1837 (Biihler 1987).
Wilhelm was a physicist, insofar as any polymath of the day can
be so labelled, and his brother Ernst Heinrich is the physiologist
who started Fechner on the path to posterity. So the mathemat-
ics available to Fechner came from Gauss through the Webers,
and uses directly the estimation methods we call least-squares.
But at what price? The Fechnerian approach to psychophysical
scaling is outside real time, a mistake that Herbart didn't make.
The computations are now tractable, and the implicit model
matches some of the invariances of observable data.
Fechner weakened the theory seriously in order to be able to
match some data properties to theory. In doing so he gave an
impetus to psychology as an experimental activity, for which we
should be eternally grateful, but not as a science in the way
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 141
Commentary/Murray. History of psychophysics
Herbart, without any data to hand except his own introspec-
tions, had tried to do - until he backed off, leaving it, as he said,
to better mathematicians.
Actually, both Fechner's and Plateau's laws are special cases
of a generic differential equation used by some Russian writers,
but as they are both wrong as models of real data (precisely
because they are not in real time and are invertible) we can
forget them both. Biological systems are what is called dissipa-
tive (and hence irreversible in time); the classical mathematics of
conservative physical systems, which was all that was available
to Fechner and later to Wundt, doesn't apply (West 1985;
Wiggins 1990). But dynamical systems have their own distinc-
tive properties. The ones of interest to psychophysics are: that a
system with innumerable internal elements does not have
infinite dimensionality and that nonlinear dissipative systems
are in real time, mathematically and empirically (Gregson
1988). In other words, nonlinear systems are good candidates for
discovering what our data look like when we stop filtering out
their information with Gaussian assumptions. The reason the
Gaussian assumption of normally distributed residual noise
distribution is wrong is not just the familiar one, that the mean
and variance are not sufficient statistics for psychophysical data,
but also that a Gaussian process in real time is the product of an
infinite dimensional attractor and we have not got infinite
dimensionality. One should also note that linearity is not com-
patible with distributions having long tails. We often have
unidentified variance, which we fudge by treating it in its
representation as noise, but that is ignorance and not science.
Murray uses examples from vision to show how psychophysi-
cal equations become inadequate, either as data representations
or as analytically tractable operations, when we move from
unidimensional to multidimensional cases. He doesn't quite
push this point home mathematically, so it is worth emphasis-
ing. In the case of vision, with white, grey, and surrounds, SI,
S2, S3, the observer has to process, either in parallel or in rapid
sequence, three inputs each of which could have been pre-
sented separately as the sole focus of a psychophysical modelling
exercise. But multidimensional psychophysics, with intercon-
nections between copresented inputs, does not follow just by
glueing Fechnerian or Delboeufian or Plateaunian equations
together and hoping that the input or output terms of one
equation become the threshold or correction parameters of the
others. Obviously, as symbol manipulation in uninterpreted
equations, you can do that, but it does have to mean something
as well. One has to write axioms for how dimensions interact
(Gregson 1992), and one has to try to separate very particular
peripheral mechanisms located in the psychophysiology of one
modality (such as vision) from some generalized interactions at a
higher level. It can be a fatal mistake to argue from one highly
specialised modality to psychophysical generalizations which
should a priori be as modality-neutral as the unidimensional
equations Murray reviews.
One could say that the two great errors which have impeded
the development of psychology as a science are the belief that
one cannot measure psychological qualities (which Murray
traces out nicely) and the adoption of statistical inference while
ignoring inverse probability, due to the conceptual obtuseness
of Venn about the same time as von Kries (Dale 1991). It has
taken psychology over a century to get back on track.
142 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
The discovery of the psychophysical power
law by Tobias Mayer in 1754 and the
psychophysical hyperbolic law
by Ewald Her ing in 1874
Otto-Joachim Grusser
Department of Physiology, Freie Universita't, Berlin 33, Germany;
Department of Comparative Physiology, Rijksuniversiteit, Utrecht,
Netherlands
Electronic mail: gruesser@zedat.fu-berlin.dbp.de
When Plateau (1872a; 1873) described the psychophysical
power law he was not aware of a predecessor who had used the
same equation to relate stimulus parameters and visual percep-
tion: Tobias Mayer (1723-1762), professor of applied mathema-
tics (Okonomie) and director of the astronomical observatory at
the University of Gottingen (1750-1762), one of the leading
cartographers and astronomers of his time (Forbes 1980; Kaest-
ner 1762). His interest in visual psychophysics originated from
practical problems arising in his scientific studies, although his
curiosity about this field and the optics of the eye can be traced
back to his high-school days at the Latin School ("Lyceum") in
the southern German city of Esslingen. As a fellow of the
"Alumnat," he later had to teach younger pupils, a duty which
led to Mayer's first mathematical book, published at the age of
18 (Mayer 1741). In his second book on applied mathematics
(Mathematischer Atlas, 1745), he considered the optics of the
eye as part of applied mathematics and gave an outline of a
quantitative colour theory, which later led to the design of a
three-dimensional (hexaedric) colour space composed of 819
different colours as defined by the mixture of three colour
pigments. Using the techniques of his time, Mayer attempted
what was later done more precisely by Ostwald and Munsell
(Krantz 1972) and invented a metrical colour system with a
defined position for each hue of all possible surface colours
(Mayer 1758; 1775; for details see Grusser 1989).
In reading off the scales of his astronomical instruments
Mayer became aware that the precision could be improved by
repeated measurements; he designed for this purpose a "repeat-
ing circle" through which the telescope was moved to determine
the position of a star several times (Mayer 1754a, cf. plate 13 in
Forbes 1980). Mayer also recognized, however, that the preci-
sion of micrometer readings depended on the illumination.
Most likely he did not know that as early as 1674 the English
astronomer and physicist Robert Hooke (1635-1703) had dem-
onstrated at a Royal Society meeting that visual acuity could be
measured by means of black and white gratings. The report on
Hooke's demonstration was first published in 1756 by Birch.
Hooke believed visual acuity to be the result of the "diameter of
the optic nerve fiber forming the inner layer of the retina"
(Hooke 1705, p. 98). Subsequent to this demonstration, it was
repeatedly mentioned in the scientific literature that visual
acuity depends on the strength of illumination (e.g., Berkeley
1709; Wolff 1725), but no quantitative investigation of this
relationship was made.
Mayer eventually became curious about this dependence and
performed psychophysical experiments on visual acuity in 1753
and 1754. He presented his experimental data and conclusions
at a meeting of the Societas Regiae Scientiarium at the Univer-
sity of Gottingen on April 6th, 1754 (Mayer 1754b), and pub-
lished a full report one year later in the society's journal,
Commentarii Societatis Regiae Scientiarium Gottingensis, vol.
4 (Mayer 1755; English translation, Scheerer 1987). To measure
how visual acuity changed with the illumination of the stimulus
pattern he first used dots of different sizes, but soon became
aware that applying gratings or checkerboards was more suit-
able for determining the dependency of the minimum visibile on
the illumination strength of the stimulus patterns (Figure 1). To
illuminate the pattern he used a candle and varied the degree of
illumination by changing the distance d of the candle from the

Tobias Mayer 1755 V - a " 1 = k. I . 0 1 6 8
0.030- checkerboard 1. k - 1/73
checkerboard 2. k 1/99
a vertical gratings 1:1, k 1/79
0 vertical grating] 1:3. k 1/52
\ 0.020-
"o
0.015
.- 0.010
5 0.008
.a"
*" 0.005
0.01 0.1 1 10
I, relative stimulus intensity (illumination of pattern)
Figure 1 (Griisser). Relationship of visual acuity (ordinate) in
seconds of arc (left side) or minutes of arc (right side) plotted as a
function of the illumination of the stimulus pattern (abscissa).
The data published in Mayer's 1755 tables are plotted in a
logarithmic coordinate system. The straight lines correspond to
the power function described by Mayer for the respective data.
Inset: examples of the stimulus patterns used by Mayer (from
Griisser 1989).
stimulus pattern. To determine the minimum visibile a, the
distance between observer and stimulus pattern was altered
until the observer could no longer discriminate the gratings or
the checkerboard pattern from a homogeneous grey. The results
of his studies, which Mayer performed "at night" with the
"direct light of a tallow candle" were presented in several tables.
The numbers in these tables were presumably the algebraic
means of repeated measurements. For the relationship between
the minimum visibile a and the distance d, Mayer found the
following rule:
a = k . dU3 [seconds of arc]
whereby the constant k varied for the different gratings and
checkerboard patterns used (Figure 1). Since the light intensity
/ decreased with the square of the distance d between light
source and visual pattern, Mayer rewrote Equation 1:
a = k . 7~"6 [seconds of arc]
Visual acuity V is defined as the reciprocal value of the minimum
visibile a. Thus Equation 2 can be transformed into
V = I/a = k* . Z"6 [seconds of arc"1]
In Figure 1 the data from Mayer's tables are plotted on a log-
coordinate system and the power functions (Equation 3) as
published by Mayer are drawn as straight lines. It is evident that
the relationship between empirical data and Mayer's power law
is very satisfactory. From Equation 3 Mayer immediately de-
duced that a human observer may be used as a subjective
photometer. He measured visual acuity for daylight conditions
and applied Equation 3 to determine daylight illumination:
Moderate daylight corresponded in his measurement to approx-
imately 25 candles lighting the stimulus pattern from a distance
of one foot.
Without much discussion of philosophical or epistemological
questions, Tobias Mayer used quantitative rules for the perfor-
mance of a human observer in psychophysical studies. His
application of mathematical rules to subjective sensations was
obviously done in the same spirit of enlightenment which
characterized his main scientific work in astronomy, mathema-
tics, and cartography (Forbes 1971; 1972; 1980; Lichtenberg
Commentary /Murray: History of psychophysics
1775). A mechanistic interpretation of the sensation, percep-
tion, and actions of man was a characteristic feature of many
scientists living during the period of European Enlightenment,
best expressed by the book L'homme machine (1748), published
5
by J. O. LaMettrie (1709-1751). The physiologists and anato-
mists of his time, however, did not grasp the impact on visual
5 psychophysics of Mayer's discoveries. Even Albrecht von Hal-
0.8 i ler (1708-1777), who was professor of physiology and anatomy at
1 the University of Gottingen from 1736 to 1753 (Balmer 1977)
and was personally acquainted with Mayer, did not mention
0.5 _ Mayer's law in his monumental Elementa physiologiae corporis
i humani (1759; 1768). He did cite Mayer's publications on visual
0.4 > acuity, however, within a long list of different reports on the
0.3 minimum visibile as published by Hooke, Buffon, Musschen-
broeck, Reaumure, and others. Haller, like Hooke (1705) and
Wolff (1725) before him, considered the "size of an optic nerve
0.2 fiber" within the retina to be a limiting factor of visual acuity, but
he did not mention the functional relationship between lumi-
nance and visual acuity. Mayer's power law is certainly un-
known to most modern sensory physiologists, but it was still
mentioned by Hueck(1840), Weber (1852), Fechner (I860), von
Helmholtz(1866), K6nig(1897), and Hecht (1928). Lichtenberg
discussed Mayer's psychophysical discoveries in his lectures on
physics (Gamau 1811).
In one of his unpublished manuscripts extant at the Gottingen
University library (Mayer 13, dated January 22nd, 1759) Mayer
protocolled experiments on the problem of how to define
quantitatively the degree of shading of coloured surfaces by
adding either white or the "colour of shadow' (a mixture of red,
yellow, and blue pigments) to the hue to be shaded. He became
aware that visual perception along the grey tone scale is a
nonlinear function of the amount of "colour of shadow" added,
but he did not go into quantitative details on this problem. As
another note and a drawing made in another manuscript (Mayer
15i3, On colours) indicate, Mayer also tried to define the three-
dimensional illusion of shaded drawings of objects quantita-
tively, that is, he was aware of the problem of "shape by
shading."
When Ewald Hering (1834-1918), professor of physiology at
Vienna, Prague, and Leipzig, entered the field of visual psycho-
(1) physics in the early 1860s, he investigated predominantly prob-
lems of binocular vision for about 10 years (Hering 1931,
publications 25-36). In the early 1870s his main interest shifted
to the physiological mechanisms underlying the perception of
the grey tone scale, simultaneous contrast and colours including
colour blindness, topics which characterized Hering's publica-
(2) tions on vision up to 1915 (papers 37-84 in Hering 1931). Hering
was soon dissatisfied with the logarithmic law of Fechner (1860)
and not only rejected Fechner's "metaphysical idea' that the
psychophysical parallelism between brain events and percep-
(3) tual phenomena is characterized by a logarithmic relationship,
but stated that the nonlinear transformation occurs in the sense
organs or at an early stage of neuronal data processing. Hering's
assumption of a linear correlation between perceptual events
and brain activity was similar to that of Mach (1868). Hering also
questioned whether Fechner's Law, deduced from Weber's
early observations (Weber 1852), was a sufficient phenome-
nological description of psychophysics. He found considerable
deviations from Weber's rule (A/// = const) when measuring
brightness sensation along a white-black, grey-tone scale. He
found that plotting perceived brightness as a function of the
logarithm of stimulus intensity results in an s-shaped curve
(published in Hering 1920) and believed that such, a nonlinear
compression function was due to the interaction of two antago-
nistic processes. Hering called these processes "assimilation"
and "dissimilation" and presumed that they characterized the
function of neuronal networks. "Excitation" and "inhibition" are
the corresponding terms in modern neurophysiology. For the
perception of brightness and darkness Hering assumed that a
dissimilative process W (white) and an assimilative process S
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 143
Commentary /Murray: History of psychophysics
(black) interact with each other, leading to the following equa-
tion for the sensation of a "whiteness' (Hering 1874; 1920):
E = W/(S + W) [units of sensation] (4)
When only the perceived brightness of a spot of light is consid-
ered, Hering assumed that Equation 4 could be transformed
into a function relating the brightness of the light and the
stimulus luminance /, which would result in the notation used in
Murray's target article as:
E = a//(l + kl) [units of sensation] (5)
with a and it as constants. This hyperbolic function indeed gives
an s-shaped curve when / is plotted on a logarithmic scale
(Figure 2; Hering 1920). Equation 5 is identical to the formula-
tion of the Michaelis-Menten kinetics and the equation applied
by Naka and Rushton to describe the horizontal cell responses of
the fish retina as a function of the stimulus luminance /. Naka
and Rushton (1974) also demonstrated that for certain stimulus
conditions an exponent a < 1 has to be added to describe the
experimental data reliably (cf. also Foerster et al. 1977 for cat
horizontal cells):
E = al" I (1 + kl") [mV] (6)
Hering (1920) discussed extensively how Equations 4 and 5,
which should be called Hering's Law (Grusser 1979), depend on
different levels of adaptation and discovered that simultaneous
contrast phenomena also follow the same rule. For the spatial
summation in retinal ganglion cells an analogous rule was
described (Buttner & Grusser 1968; Grusser et al. 1970), which
was generally called "shunting inhibition" in models of neuronal
networks like those applied by Reichardt and others (cf. Furman
1965).
In summary, when one considers the history of early endeav-
- 3 - 2 - 1 0 1 2 3
log 10 stimulus luminance [arbitrary units]
Figure 2 (Grusser). Relationship between the perceived
brightness of a spot of light seen on a grey background (ordinate)
as a function of stimulus luminance (abscissa, arbitrary log-
arithmic scale). Hering plotted these curves corresponding to
Equations 4 and 5. The two different curves relate to different
brightness functions at different background luminances lead-
ing to two different "Eigenfarben" of the background (0.5 and
0.3 on the black-white scale of the ordinate). On the abscissa a
base-ten logarithmic scale is added to the original figure (from
Hering 1920, Figure 17).
144 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
ours to relate the phenomena of visual sensation and perception
to stimulus intensity and to deduce quantitative rules ("laws")
from the experimental data, one is led to a revised view of the
discoveries of such laws. The power law of psychophysics,
usually attributed to Plateau or Stevens, was first described
more than two generations before Plateau by Tobias Mayer in
1754/55. The same is true for the hyperbolic function, which
relates stimulus intensity to perceived brightness (or darkness)
and corresponds to the Michaelis-Menten equation. This non-
linear function was first applied to modem neurophysiological
data by Naka and Rushton in 1966. The law itself was discovered
by Ewald Hering to interpret his observations in visual psycho-
physics three generations before its application in modern
neuroscience.
Psychophysics, its history and ontology
Horst Gundlach
Institut tar Geschichte der Neueren Psychologie, UniversitSt Passau, 8390
Passau, Germany
Psychophysics started out as one individual's invention. It
changed into a scientific community enterprise nearly two
decades after its cardinal components were elaborated. It is
therefore appropriate that attention turns back to Fechner
whenever the muddle that is psychophysics appears to need
elucidation. Murray's initiation of a fresh look at the early stages
of psychophysics is, in my opinion, praiseworthy and promising
for present-day psychophysics. I assume, however, that his
"perspective for viewing" needs the support of a few more facts,
and that his observations on contemporary psychophysics could
profit from a critique cutting somewhat deeper.
Murray's historical narrative shows the characteristic of most
narratives of psychophysics - solid anachronism. That is, early
stages are accounted for from the perspective of later concerns.
Yet Murray fares better than most, as he directs attention to the
often-neglected "inner psychophysics." He misses, however,
the true nature of the relationship between inner and outer
psychophysics.
When Fechner had hisflashon Fechner Day 1850, he had in
mind nothing but what he afterwards called inner psycho-
physics. He developed outer psychophysics later as a subordi-
nate, subservient enterprise to corroborate the fundamental
inspiration.
This inspiration consisted in the conception of the universe as
an undivided, non-Cartesian entity. In technical terms, Fech-
ner espoused the metaphysics of a neutral ontological monism,
that is, while rejecting both materialism and spiritualism, he
attempted to unify their positive assumptions. The surmised
fundamental entity nevertheless exhibits two forms of appear-
ance, physical and mental, or, to use another word, spiritual. In
short, a phenomenalist dualism went with the cardinal ontologi-
cal monism. This was, of course, in Fechner's times not a new
vision of the universe, but a rather unoriginal one, usually
termed Spinozism by its opponents.
The novelty Fechner introduced was the notion that the two
appearances of an underlying primary entity were coupled in a
definite way, namely, the numerical values of counterpart
appearances behave as a logarithmic function. This implies that,
viewed from the other side, they behave as an exponential
function. In simple words, multiplication in the physical realm
is matched with addition in the mental realm. So far, this was
bare and barren a priori fabrication and reproached as such by
Wilhelm Weber.
To wit, any part of the primary entity has the faculty of double
appearance: people, animals, plants, minerals, stars, and so on,
not only brains. The mind-brain problem is not at the heart of
inner psychophysics.

The auxiliary science, outer psychophysics, created to gener-
ate empirical support for Fechner's quaint metaphysics, intro-
duces a new ingredient: causality. Inner psychophysics assumes
only the relation of identity between matching appearances.
Outer psychophysics, however, presupposes a causal chain from
outer stimulus to activity modulation at the sense organ to
nervous excitation to central activity. The brain activity caused
by stimulation is ontologically identical to sensation, and the
numerical relation between their matching values is expressed
by the basic psychophysical law. Outer psychophysics needs
identity plus causality. It assumes proportionality between
stimulus and brain values, and logarithmicity between brain and
sensation values.
This is what Fechner tells us outer psychophysics is to be
about. This tale fits his desire to substantiate the claims of inner
psychophysics. And it has been repeated over and over again, as
Murray has now repeated it. Fechner's narrative is a fable,
however, and one that fails to correspond to what he and other
psychophysicists actually do when they turn from reasoning to
experimenting. This fable has prevented generations from ask-
ing the obvious, because the answer was seemingly already
given. The obvious question is: What is the realm of reference of
outer psychophysics as a research domain? What do statements
in outer psychophysics refer to?
A negative answer first: They certainly do not refer to sensa-
tions or other mental phenomena. Consider the weight experi-
ments by E. H. Weber (1834; 1846; 1851; 1978). They are about
objects, namely, weights. They explore human capacity to judge
differences between weights under specified conditions of expo-
sure without the assistance of instruments. This is then com-
pared with judgements succoured by precision balances.
Consider the case of the so-called extensive sensations. Fech-
ner, like others, judges line lengths. He talks about the lines and
produces unaided judgements about their lengths. These he
compares with finer judgements produced by using a ruler. The
case is more complicated in sound or light perception. Here the
judgements pertain to either objects producing sounds or lights
(e.g., stars) or to the events produced. This depends on the
instruction to the subjects. Each is feasible, but in neither case
do any sensations need to be talked about. As sensations do not
belong to the realm of reference of outer psychophysics, there is
a fortiori no place for sensation strength and no necessity to
discuss whether sensations are measurable.
Empirical work in outer psychophysics, today "psycho-
physics" for short, assesses the fineness of calibration of various
senses under various circumstances in determining the values of
diverse attributes of objects, or, depending on closer ontological
analysis, events. The way out of the muddle of psychophysics is
the analysis of what is done in psychophysical experiments and
the construction of a sober ontology freed from Fechner's
strange ambition of demonstrating empirically a metaphysical
system of the universe (Gundlach 1992). Plain (outer) psycho-
physics simply refers to objects or events and humans (and
animals). There is no place or need for sensations and similar
ontologically dubious characters.
We must forget Fechner's fable, originating from the desire to
corroborate his metaphysics, that smuggles in the contraband of
sensations. Sensations may help in obtaining judgements, but
they would play a similar role in obtaining instrument-aided
judgements, for example in meter reading or scale inspection.
Whatever they might be, they are none of psychophysics's
business.
Commentary I'Murray: History of psychophysics
Inner psychophysics, neurelectric function
and perceptual theories
Stephen Handel
Department of Psychology, University of Tennessee, Knoxville, TN
37996-0900
Electronic mall: handel@utkvx.utk.edu
What is the appropriate role for psychophysics in a theory of
perception? From one point of view it is pivotal; the perception
of the complex overlapping coordinate worlds within seeing and
hearing can be and will be explained by simple psychophysical
functions. From a different point of view, psychophysics is
largely irrelevant. Seeing and hearing involve the pickup of
complex invariants within a context, and simple psychophysical
functions can have little predictive power to explain that pickup.
Although a middle ground is possible based on the ability to
abstract attributes or dimensions of a complex stimulus and
thereby explain quantitative aspects of the precept (but see
Garner 1974, and Melara & Marks 1990, for constraints on this
ability), this has proven to be difficult. I have come to believe
that the usefulness of psychophysics is limited. Any attempt to
obtain a unified law will fail because there is as yet no unified
neurelectric function as a consequence of perceptual systems
evolving to group organized structures.
Many different attributes have been scaled. For example,
Gescheider (1984, p. 230) lists 33 attributes along with their
power law exponent. These are truly a mixed bag. At one end,
there are dimensions in which the magnitudes could be signaled
by the rate of firing of a single type of receptor (e.g., brightness,
pressure on palm, and angular acceleration). At the other end,
there are dimensions in which the magnitudes could not be
signaled by any single type of receptor (e.g., visual length or
area). In the middle, there are attributes that could be signaled
by some neural property, but not necessarily the firing rate of
one type of fiber (e.g., sound duration or length of silent
interval). It is the very diversity of these attributes that suggests
there is no common psychophysical law in spite of the fact that it
is possible to fit a logarithmic or power function to all of these
attributes.
If we look more closely at the possible neurelectric functions
underlying inner psychophysics, it is clear that transformations
between intensity and firing rate at the receptor level are only
the first step in an intricate set of transformations at intermedi-
ate nuclei. The neural firing at the cortical sensory areas repre-
sents the signal by the pattern of firing across fibers. There is no
subset of fibers that faithfully represents the firing pattern at the
receptor level. Instead of considering the perception of loud-
ness that has been extensively reviewed, I would like to use
sound duration as an illustrative example. A recent review of
work on frogs (Feng et al. 1990) demonstrates the immense
transformations that take place between the auditory periphery
and the thalamus with respect to the neural representation of
natural sounds such as mating calls. At the periphery, regardless
of the temporal characteristics of the sound, all auditory fibers
will respond to any signal as long as there is sufficient energy
within that fiber's excitatory frequency range. Nearly all fibers
increase their firing rate to longer signals (i.e., high pass).
However, at higher levels, neurons become more sensitive to
temporal characteristics of the signal. In contrast to peripheral
fibers, the majority of neurons in the thalamus respond maxi-
mally to signals below a certain duration (i.e., low pass) or to
signals within a certain range of durations (i.e., band pass). In
addition to those neurons coding duration, there are other
groups of neurons that code amplitude modulation irrespective
of amplitude. The discrimination of duration and modulation
becomes based on the patterning of the neurons. The firing
pattern does not mimic the time-intensity pattern of the periph-
eral fibers.
Each sensory system is organized into parallel information
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 145
Commentary /Murray: History of psychophysics
channels for different spectral and temporal attributes. The
perception of real events is therefore based on the firing pattern
both within and across channels. This is a common strategy for
all sensory systems of vertebrates. Each channel can be innately
tuned for specific values that may signify important environ-
mental events such as conspecific mating calls for the frog and/or
each channel can be experientially tuned by the naturally
occurring environmental events (Kaas 1982; Merzenich et al.
1984), as might be the case for speech (Kuhl et al. 1992).
What this means for Fechner's concept of inner psycho-
physics is that the neurelectric function must be derivative. As
Fechner himself believed, in no sense is it a direct measure of
peripheral sensory function. The parallel sensory channels pro-
vide the mechanism for the abstraction and judgment of differ-
ent sensory attributes, but that very process of abstraction
would appear to change the neural code so radically that a
straightforward neurelectric function is impossible. What it also
means is that the typical experimental situation is unrepresenta-
tive of the natural functioning of perceptual systems (Gibson
1979). Sinusoidal tones presented in isolation are not a "pure"
signal; the tones are presented in a context, but that context is so
fundamentally unrepresentative that the resulting processing
may be distorted. It may be for this reason that many instances
of categorical perception where Weber's Law apparently is
violated (Pastore 1987) are found with complex time-varying
stimuli (Rosen & Howell 1987).
From this perspective, it is remarkable that the results from
psychophysical experiments are so consistent. Although the
concepts inherent in signal detection theory can be used with
success in physiological investigations of discrimination based
on firing rate (Kim et al. 1990), the basic issue seems to me to be
the way discrimination emerges out of firing patterns across the
neurons. The question is not how to dig the signal out of a noisy
background or whether a logarithmic or power law is more
correct. The similarities in neural organizations for abstracting
different attributes coupled with the commonalities in psycho-
physical functions relating intensity to sensation strength for
each attribute argue strongly that there is common neurological
strategy. This strategy in turn yields a common neurelectric
function. The objective is to understand the nature of that
strategy and resulting function; then I think it will be possible to
ascertain the relevance of Fechner's inner and outer psycho-
physics for theories of perception.
Fechner's impact for measurement theory
Michael Heidelberger
Philosophisches Seminar, University Freiburg, D-7800 Freiburg, Germany
Electronic mall: engelhar@sun1.ruf.uni-freiburg.de
None of the presentations of Fechner's original conception of
the psychophysical law known to me (and Murray's presentation
is no exception) give an explanation of the fact that Fechner
called his function a "measurement formula" (Massformet) and
not a "law." This might seem an unimportant detail without any
further relevance at all. My claim, however, is that the answer to
this innocent question reveals that Fechner's reasoning is based
on a highly original conception of measurement in general; one
that is at variance both with the predominant conception of
measurement in Fechner's own time and with the most popular
conception of today, the representational/operational theory of
Suppes and Stevens (Suppes & Zinnes 1963; Stevens 1946; on
the history of this conception see Newman 1974; Michell 1986).
I think that Murray's historical treatment of the psychophysical
law must be supplemented by this measurement-theoretical
dimension before we can fully understand the force and scope of
Fechner's original ideas in the line of research leading to
modern measurement theory.
146 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
Fechner's correlational theory of measurement. Fechner
(1858; 1860, vol. 2, Ch. 7; 1882, Ch. 26; 1887; for a more detailed
description see Heidelberger 1986; 1993a, Ch. 5) argued that
the measurement of any attribute Q generally presupposes a
second, directly observable attribute R and a measurement
apparatus A that can represent variable values of R in correlation
to values of Q. The correlation is such that when the states of A
are arranged in the order of Q they are also arranged in the order
of R. The different values of R are defined by an intersubjective,
determinate, and repeatable calibration of A. They do not have
to be measured on their part. The function that describes the
correlation between Q and R relative to A (underlying the
measurement of Q by R in A) is precisely what Fechner called
the measurement formula. Normally, we try to construct (or
find) a measurement apparatus which realizes a 1:1 correlation
between the values of Q and the values of R so that we can take
the value of R as a direct representation of the value of Q.
In developing his principles of sensory measurement Fech-
ner several times referred to time measurement as an analogy
(Fechner 1858, p. 2; 1887, p. 217). His argument can be
reconstructed as follows: The attribute "time' is measured by
the attribute "distance" as realized in a clock or in the movement
of the sun. Time values cannot be measured abstractly without
using their relation to distance values. Normally, clocks are
constructed so that we obtain an invertible 1:1 map from the set
of time values {t} that can be possessed by Q into the set of
distance values {d} that can be possessed by R. If we know the
value of some d (the distance covered by a clock's hand) we can
pick out directly the value of the associated t (the time value). So
in the case of time measurement, the measurement formula that
underlies the construction of a clock is a linear function d = kt,
for a constant k. Note that this measurement formula does not
state that a certain empirical law holds between time and
distance. Rather, it describes how time and distance are related
to each other by the clock that was chosen for time measure-
ment.
Consider now Fechner's transposition of this reasoning
(Fechner calls it his "measurement principle") to the problem of
measuring sensation. To measure sensations we first have to find
an instrument for which we know the measurement formula
involved. Suppose that S refers to sensation strength. The only
attribute from which we know that it varies with S is the physical
intensity / of the stimulus. The only device which could possibly
serve as a measurement apparatus relating variable values of/ to
values of S is the living human body. We know from experience
that sensation grows as a function of stimulus intensity such that
S is isotonically related to / for some specified range. (That is, for
all pairs sm, sn of sensation values there exist correlated stimulus
intensity values im, in so that: if sm is greater than s n , then im is
greater than tn.)
It is obvious that the measurement formula realized by the
human body as a measurement device is much more compli-
cated than in the case of a constructed clock. So the first problem
to be solved in order to measure sensation is to find out the
measurement formula by which the outer stimulus is correlated
with the inner sensation.
For the deduction of his logarithmic measurement formula
Fechner proceeded as described by Murray in sect. 1.1. There
are three empirical cornerstones on which Fechner builds his
argument: (1) the existence of the absolute threshold, (2) the
equality of all jnds, and (3) the "fundamental formula" (Murray's
Equation 2), as deduced from Weber's Law.
So what we call the "psychophysical law" today was first of all a
"measurement formula" for Fechner. Although any measure-
ment formula has important empirical import it does not have
the status of a law of nature. It is more akin to a definition or to
what philosophers call an analytic statement. A natural law,
however, is a synthetic, factual statement about the way vari-
ables that are measured independently of the law itself are
related to each other. For each of the variables in a law there

must exist a measurement apparatus with a measurement for-
mula before a law can be established and tested.
Stanley S. Stevens. Fechner's correlational theory of measure-
ment is important for assessing Stevens's reasoning. From a
Fechnerian point of view, Stevens's function is either a law or a
measurement formula. It could only be called a law in the strict
sense of the word if the measurement formula were known by
which sensor)' intensity is related to the subject's direct estima-
tion of it. One could interpret a finding by Wendt (1982) as such
a formula. Wendt has suggested that the subjective magnitude
of numbers is a logarithmic function of actual numbers. If this is
true, any power function relating estimated numbers to stim-
ulus intensity would imply that Fechner's logarithmic measure-
ment formula does hold between S and / after all. If one does not
ask for a measurement formula as an underpinning for Stevens's
Law one takes it for granted that there is a 1:1 correlation
between sensory intensity and its estimation. In the eyes of
Fechner, this assumption would be more a problem to be
explained than a solution.
Stevens's Law could also be taken as a measurement formula
itself. For Fechner, this would have been quite legitimate as
long as some empirically more valid alternative to the funda-
mental formula from which he deduced his logarithmic function
could be established in experiments. From Fechner's point of
view such an alternative can hope to be successful if it is
obtained by some scaling method that gives us reliable knowl-
edge about subjective equality of certain difference sensations
across the scale.
Johannes von Krles. Von Kries's intention in the article cited
by Murray (in sect. 1.5.3) was to delimit the conditions under
which an attribute can legitimately be said to be measurable
(von Kries 1882). He started from the observation that for an
attribute Q to be measurable there must be an objective and
clear sense of what it means that one value of Q is equal to
another one. If such a criterion of equality is available, any value
of Q can be compared with a standard unit of Q and the units can
be counted. In the case of time and space, said von Kries, it is
uncontroversial what equality means and how it is established.
For mass, it is conventionally fixed that two objects are equal if
they balance each other on a pan balance with arms of equal
lengths.
How is equality established for other attributes? In the
physical sciences this is done according to von Kries by forming
units of these attributes as combinations of units of space, time,
or mass. The way this combination is defined is arbitrary. The
only criterion involved is pragmatic convenience in the light of
certain empirical facts. As a result, all physical measurements
are ultimately based on comparisons of lengths, durations, and
masses. These comparisons guarantee the objectivity of physics.
In psychology this method is not viable because sensation
cannot be defined physically; but if an objective definition is
impossible then the subjective equality of sensations cannot
have an objective validity. It is of course possible to stipulate
that all jnds are equal, but this is not enough to secure objec-
tivity and would turn Fechner's fundamental formula into an
awkward convention about sensation talk, without any empirical
import. One could equally stipulate that two sensation incre-
ments are equal if they correspond to equal increments of
physical intensity. One expression would be as legitimate as the
other one. Von Kries thus came to the conclusion that Fechner's
formula and the whole idea of sensation measurement just do
not make any sense at all.
Von Kries also analysed why so many people think it makes
perfect sense to determine two sensations as equal or unequal.
They automatically transfer the objective properties of the
stimulus into the sensation and take the objective equality of two
stimuli as the equality of two sensations. (This is what Boring
later called the "stimulus error.")
It is interesting to note that von Kries's denial that sensory
intensity can be measured led to the development of the logical
Commentary/Murray: History of psychophysics
interpretation of probability a few years later. Von Kries argued
that the reasons for the nonmeasurability of sensory concepts
apply also to probability if it is taken as a psychological state. It is
impossible, indeed senseless, to measure the degree of subjec-
tive certainty with which an individual holds, expectations and
beliefs (von Kries 1886/1927, p. 3). Instead, probability has to
be conceived as a logical concept. Wittgenstein later took up this
idea, which in the hands of Waismann (1930-1931) and Carnap
(1945) became the logical interpretation of probability as we
know it today.
Von Kries was a physiologist (in Freiburg, by the way) of the
Helmholtz school but he also contributed heavily to neo-
Kantian philosophy. In their condemnation of Fechner's theory
the neo-Kantians who dominated philosophy in the last quarter
of the nineteenth century and the biophysics movement of
Helmholtz and his friends formed a strong alliance. Although
they had different motives, both schools rejected any naturalism
or physicalism in regard to the mental and saw in Fechner's
theory a relapse into Naturphilosophie of the Schelling brand,
where physical and mental categories are illegitimately con-
founded. Although Helmholtz did not explicitly mention
psychological measurement, his famous article on "Counting
and measuring" has to be seen in this context (Helmholtz
1887/1921). To von Kries's conditions for the measurement of an
attribute Helmholtz added the requirement that there must
exist an empirical operation that combines any two objects
exhibiting the attribute into a composite object that also exhibits
the attribute. The attribute can be measured if the operation is
mirrored by the mathematical laws of additivity.
Ernst Mac/). Ernst Mach opposed both neo-Kantianism and
the mechanistic outlook of the Helmholtz school and was
strongly influenced by Fechner in many different respects.
Mach responded to objections such as those of von Kries and
others that it is a delusion to think one could do without
sensations in science. Any establishment of an objective equal-
ity in science by way of comparing lengths, durations, and
masses must ultimately be based on sensation because it needs
the reading (or at least the gauging) of a material device by an
observer (Mach 1886/1922, p. 281). Mach not only defended
Fechner's measurement theory, he radicalised it and extended
it into physics.
This seems to be at variance with Mach's claim, as cited by
Murray, that "there can be no question of an actual measure-
ment of the sensations" (Mach 1886/1922, p. 67; see Murray,
sect. 1.5.3). This very citation was also pointed to by Stevens,
who wanted to show with it that opposition to Fechner was
widespread among reasonable people (Stevens 1975, p. 59). The
context of the citation reveals, however, that what Mach wanted
to say was that (as we would put it today) sensations are not
measurable on a ratio scale but on an interval scale only.
According to Mach, a psychophysical measurement formula
assigns numbers to sensations in the same way a thermometer
assigns the temperature to a state of heat.
In the above quotation, Mach uses "measurement" in the
sense of "measurement on a ratio scale." Why this strange use?
This was the predominant way of speaking about measurement
during Fechner's and Mach's time. Measurement on nominal
and ordinal scales was not seen as measurement at all. What we
today conceive of as measurement on an interval scale was seen
as a special case of measurement on a ratio scale. Take the
measurement of temperature by a mercury thermometer as an
example. Current opinion held that by using a thermometer one
is actually measuring not the real temperature of an object but
the reaction of the mercury column to the temperature of the
object. Thus, temperature measurement by a thermometer is
measurement of a certain length on a ratio scale. This was also
the opinion of Helmholtz, who described the measurement on
an interval scale as measurement of "coefficients" because the
scale is not really based on the attribute that is supposed to be
measured but on additive properties of special material objects
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 147
Commentary/Murray: History of psychophysics
such as the mercury column (Helmholtz 1887/1922, p. 93).
The spell of this reasoning was broken by the analysis Mach
gave of temperature in his Theory of heat (Mach 1896/1919, pp.
39-57). Brian Ellis called Mach's analyses of mass and tempera-
ture "the first important advances in the logic of measurement
since pre-Socratic times" (Ellis 1967, p. 242). Mach came to the
conclusion that the assignment of temperature numbers is no
"measurement" in the strict sense (i.e., a makeshift measure-
ment on a ratio scale) but an "inventory" of states of heat "by
numerical means" (i.e., measurement on an interval scale). It is
a familiar problem in the history of science that exponents of
revolutionary achievements have a hard time with their lan-
guage and are often forced to express themselves in a quite
conservative terminology that conceals their attainments from a
superficial view. This is exactly the way the quotation by Mach
should be construed.
Although Mach defended Fechner's measurement theory he
did disagree with Fechner in one very important respect that
made him his adversary in the eyes of many contemporaries. He
joined his friend Hering (who had been a student of Fechner in
Leipzig) in claiming that neither the fundamental formula nor
the measurement formula is valid for inner psychophysics. Yet it
had been the ultimate aim of Fechner to demonstrate just that!
For Fechner, the establishment of a measure of sensations for
outer psychophysics was only a preparatory if important step in
the development of "real" psychophysics. Psychophysics would
only be able to leave its provisional phase and use the developed
measure for actual measurements if it could be shown that the
logarithmic measurement formula is strictly valid for the "point
of contact" between neural excitation and sensation. If this were
the case then psychophysics could be an autonomous science
and would develop a progressive research program of its own
dealing with questions about how higher mental functions like
memory, attention, and self-consciousness are represented psy-
chophysically; if not, psychophysics would generate and shrink
to an unimportant and tiny appendix of physiology (Fechner
1882, p. 262).
Yet Mach and Hering and many others came to the conclusion
that the transformation between mind and brain is proportional,
that is, linear (Hering 1875, p. 330; Mach 1868, p. 12). They did
not shrug off Fechner's contrary opinion on the grounds that it
was too metaphysical, as Murray claims (sect. 1.2). They
thought that if a logarithmic law were valid for inner psycho-
physics we would be unable to perceive our environment
adequately (Hering 1875, p. 321). Hering and Mach also held
that the logarithmic character of Fechner's formula is due to the
physiological transformation between peripheral and neural
excitation. Yet they were both convinced that a linear relation
between mind and body would be much more consistent with
and serviceable to Fechner's philosophy of psychophysical par-
allelism than a nonlinear one (Hering 1875, p. 310; Mach
1886/1922, pp. 50-51, 305). A linear relation would guarantee
that any property of the mental exactly matched to one property
of the physical.
The significance of Inner psychophysicsfor Fechner. Fechner,
however, abhorred a linear law because this would have meant
denying the absolute threshold for inner psychophysics. He
believed that not only his special version of psychophysics but
any future psychophysics in general would be impossible with-
out the concept of an inner threshold. For Fechner, giving up
the threshold would also have meant giving up the psychophysi-
cal representations of the relation between sleeping and waking
and between consciousness and the unconscious that were so
dear to his heart (Fechner 1877, pp. 70-71; 1882, pp. 255-63).
The denial of the threshold, Fechner feared, would ultimately
lead to some kind of eliminative materialism or some kind of
reductive identification of the mental with its physical correlate;
there would be no longer any necessity for science to consider
the sensory as an autonomous realm.
Fechner's hope was that the development of an advanced
148 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
inner psychophysics in his sense would ultimately give evidence
for his functionalist view of the relation of mind and body (see
Heidelberger 1988; 1993b). Long before Putnam (1967/1985)
and Fodor (1968), Fechner held it theoretically possible that
mental states could be realized by systems other than the brain.
He could of course not yet use the computer analogy, but he
frequently referred to musical instruments as a simile. One and
the same music can be played on different kinds of instruments.
In fact, he thought that death is the event when (to use
contemporary wording) the "program" realized by our brain will
be taken over by another "hardware": by the causal chains that
were initiated by our actions during our lifetime. This is not
more speculative than similar fantasies by AI adherents like
Hans Moravec (1988, Ch. 4). Fechner was convinced that from
his brand of functionalism one can deduce that mental and
physical properties are different aspects of one and the same
substance. The logarithmic character of the law for inner psy-
chophysics would thus make it impossible to (type-)identify the
conscious with the material or to eliminate it. Fechner's philoso-
phy of mind and his inner and outer psychophysics are
inseparable.
The chimera of psychological measurement
Gail A. Hornstein
Department of Psychology and Education, Mount Holyoke College, South
Hadley, MA 01075
Electronic mail: ghornste@mhc.mtholyoke.edu
On the morning of October 22, 1850, while lying in bed, Gustav
Fechner gave birth to a scientific psychology. There was no
blood, only an intuition, and the child took a logarithmic form. A
century has passed, and we still struggle to understand how
psychophysics, born of a mystic, can reveal the science of mind.
Even at the start there were problems. According to Murray,
Fechner "had little respect" for direct estimations of perceptual
magnitude, seeing them as having "no more value than ratings of
'beauty'or other subjective impressions" (sect. 1.2, para. 1). Yet
it was precisely these kinds of estimates that Stevens later
claimed were fundamental to any objective study of perception.
Ambiguity about the quantitative status of psychophysical mea-
sures persists, and is even occasionally blamed on subjects, who
apparently "give ratio judgments even when the task is devised
to yield difference judgments" (sect. 1.3.3). There also seem to
be circumstances (giving rise to the phenomenon of "categorical
perception") where subjects are able to discriminate only a few
of the possible values along a dimension of stimulus intensity
(sect. 1.4.1).
The difficulty, of course, is that psychophysical data are
judgments, not automatized read-outs of some underlying
quantitative "reality." One might suppose that psychophysicists
- whose whole enterprise is founded on an assumed difference
between the physical and the psychological worlds - would be
willing to recognize the inherent subjectivity of perceptual
judgments. Of course we can impose numbers on these or any
other forms of judgment, but doing so in no way mitigates the
ambiguity that judgment entails.
Even such basic psychophysical ideas as "threshold" raise
fundamental conceptual problems that have yet to be resolved.
To assume a threshold is to impose a qualitative discontinuity
between two adjacent points on a continuum. Such an assump-
tion casts doubt on the possibility of continuity between points
and makes the construction of scales where points are equidis-
tant from one another especially problematic.
The history of psychophysics is the story of how psychologists,
unable to escape such quandaries, chose to act as if they weren't
there (see Hornstein 1988 for a fuller account). Murray's title
suggests that he intends to examine this history; instead, he

largely reenacts it. This is most apparent in section 1.5, where
Murray (in the tradition of many of his psychophysicist col-
leagues) carefully takes note of von Kries's original critique -
which discounted even the possibility of psychophysical mea-
surement - only to act in subsequent sections of the article as if
this argument had no force. But force it has, even after a
hundred years, and Murray's dismissal of this critique, known in
the nineteenth century as "the quantity objection," reveals a
surprising presentism in his approach.
Von Kries argued that numerical values assigned in psycho-
physical judgments are not quantitative. Rather, they are labels
for qualitative distinctions, arrayed on what Stevens later called
ordinal scales. Just as relative degrees of loudness can be
differentiated by terms like "noisy," "piercing," and "deafen-
ing," we can call tones "2," "5," and "9" on a scale of loudness.
But such judgments are subjective and qualitative, and von
Kries felt that using numbers as labels gave a misleading sense of
precision where none really existed.
According to Murray, Fechner dismissed von Kries's argu-
ment because to have taken it seriously would have meant
abandoning psychophysics and returning to "a wasteland where
psychology had no scientific status and no body of measure-
ments to contemplate" (sect. 1.5.3, para. 2). Precisely. A field
which justified its existence on the grounds that it constituted
the first scientific psychology could hardly afford to doubt the
underlying quantitative foundation of its work. But there was no
way to prove von Kries wrong. (If anything, the subsequent
century of debate, with its plethora of psychophysical "laws" and
constant bickering about definitions of measurement, may have
proved him right.) Fechner's response to von Kries's critique,
while not exactly a model of scientific open-mindedness, was
nevertheless brilliant sociology. Instead of entering into an
argument about the substance of von Kries's claims, Fechner
simply pointed to the fact that subjects (in experiments such as
Delboeuf's) seemed perfectly able to perform psychophysical
tasks (sect. 1.5.3, para. 2). This fact is, of course, beside the
point: Subjects will attempt to perform any task they are given,
and their ability to choose among degrees of greyness could
easily be seen as just the sort of qualitative judgment von Kries
described.
Later psychophysicists refined Fechner's strategy. One fre-
quent tactic was to create mathematical rules that enable one
psychophysical scale to be related systematically to another.
Such transformations: (a) obscure the fact that methods that
ought to yield the same data produce different results, while (b)
making it appear as if such differences are merely mathematical
variants of one another. That the "rule" created to permute one
set of values to another is itself mathematical is also helpful,
because it serves to reinforce the basic assumption that percep-
tual processes are in fact capable of being quantified.
Attempts to triangulate perceptual phenomena by using dif-
ferent scales have the effect of masking the uncertainty that is an
unavoidable part of the psychophysical enterprise. We can
understand these attempts as analogous to those of the British
localizationists that Star (1985; 1986) analyzes. Unable to recon-
cile findings from animal experiments with data from clinical
observations of brain-damaged patients, advocates of localiza-
tion of function relied on whichever set of data best supported a
particular hypothesis. This triangulation among different lines
of work both deflected attention from the uncertainties in any
particular set of data, and made all findings seem more robust
(see Star 1989 for a fuller account).
Murray tells us to "rest content [as he simply points] to the
enormous popularity of a power law in twentieth-century re-
search following Stevens's espousal of it" (sect. 1.3.4, para. 3). In
so doing, he joins a well-established tradition in which the
existence of psychophysical research is taken as a sign of its firm
foundation. (Such a criterion might please the followers of J. B.
Rhine, who yearn to have their hundreds of ESP experiments
taken seriously as science; see Alcock: "Parapsychology: Science
Commentary/Murray: History of psychophysics
of the Anomalous or Search for the Soul" and Rao & Palmer:
"The Anomaly Called Psi: Recent Research and Criticism' BBS
10[4] 1987.) Murray, like the scores of researchers he cites,
might well ask: Why would psychophysical research continue to
be done if its assumptions are so problematic? Why, indeed?
"Fechner's whole endeavour" may have "rested on quicksand,"
as von Kries put it (sect. 1.5.3, para. 1), but psychophysicists
have not let this fact slow their labors one jnd. Even in the
1940s, when the quantitative status of psychophysical data was
the subject of formal challenge (Ferguson et al. 1940), re-
searchers persisted. Stevens, following Fechner's lead, help-
fully redefined measurement in a way that obscured its link to
quantification. If even ordinal scales qualified as a form of
measurement, how could there be any doubt about whether
psychophysical judgments were scientific? The introduction of
magnitude estimates as a "direct" way to assess a subject's
perception was also useful. With the experimenter's own sub-
jective judgments now reified as "the standard," psychophysi-
cists could enter an Alice-in-Wonderland world in which num-
bers meant whatever they wanted them to mean. Perhaps this is
why they took to calling certain measures "confusion scales"
(sect. 1.2, para. 1).
We can certainly agree with Murray that "responses gener-
ated to stimulus arrays . . . constitute a body of factual informa-
tion" (sect. 2.1, para. 4). The question is whether these re-
sponses provide any evidence to support psychophysical "laws."
Murray and his colleagues keep claiming that they do, and since
science is as much an exercise in persuasion as anything else,
these repetitions have the effect of making psychophysical data
appear robust. But if, as Murray insists, "what can be said clearly
can be said briefly" (sect. 3, para. 7), then perhaps it is time it be
said: Psychophysics is to a science of psychological measure-
ment what the Red Queen's words to Alice were to the truth.
History of psychophysics: Some
unanswered questions
Lester E. Krueger
Department of Psychology, Ohio State University, Columbus, OH
43210-1222
Electronic mail: krueger.2@osu.edu
This commentary will discuss in turn four issues raised by
Murray's History: (1) contrast effects, (2) psychophysical func-
tion, (3) the nature of variability, and (4) the physical scale.
1. Contrast effects. In order to reveal the true underlying
psychophysical function, it is necessary to pare away bias and
context effects not only at the cognitive level, but also at the
sensory or neural level (Krueger 1991; Wasserman 1991). Inhib-
itory sensory or neural interactions between stimuli are evident
in the contrast data Murray cites. He notes (sect. 2.2) that when
greys are viewed against a white background, lightness con-
stancy (which is consistent with both the power function and the
logarithmic function) is found only for medium greys. Darker
greys look darker as illumination is increased, whereas lighter
greys look lighter. Murray (sect. 1.3.2) also mentions the asym-
metrical effect of contrast: "White areas have a darkening effect
on dark areas . . . whereas dark areas may have no effect on
white areas."
According to Murray, Delboeuf's function, S = log [(c + /) /
c], can deal with these contrast effects, whereas "neither a
power law nor a logarithmic law can predict how all contrasts
will look under increasing, or decreasing, illumination" (sect.
2.2, see also sect. 1.3.3). It does not trouble me, however, that
the power and logarithmic functions cannot handle all contrast
effects, because contrast effects represent distortions that must
be gotten rid of, not clasped to one's bosom, when devising a
psychophysical law. Dual-stimulus tasks, which introduce new
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 149
Commentary/Murray: History of psychophysics
factors (such as contrast effects) into the perceptual and judg-
mental process, should be avoided as far as possible (Krueger
1989r, "Keep it simple," sect. 2.3), not given center stage, as is
the case with Murray.
2. Psychophyslcal function. Equal stimulus ratios (which de-
fine lightness constancy) may produce equal subjective differ-
ences (Fechner's logarithmic function), equal subjective ratios
(Plateau's power function), or neither (Delboeuf's function).
Delboeuf's function produces nearly equal subjective differ-
ences and nearly equal subjective ratios, however, as shown in
Murray's Table 2; and thus it represents a (potentially very
useful) compromise between a logarithmic function and a power
function. The power function and logarithmic function are
related. The power function very closely approximates the
logarithmic function when the power-function exponent ap-
proaches zero (see Krueger 1989a, Fig. 1). The Delboeuf func-
tion evidently bridges the gap between the power function and
the logarithmic function. When c = .5, as in Murray's Table 2,
the Delboeuf function better approximates the power function,
producing more nearly equal subjective ratios than subjective
differences. As c approaches zero, however, Delboeuf's func-
tion better approximates the logarithmic function, as Fechner
(1877) pointed out (target article, sect. 1.3.1).
The extra degree of freedom provided by c enables Del-
boeuf's function to obtain some of the greater expansivity found
in the power function, as compared with the logarithmic func-
tion, and thus to equal or slightly surpass the logarithmic
function in predicting Delboeuf's data (see Murray's Note 3).
Because it is estimated from the data rather than based on an
independent measure of the "natural light of the retina" (Helm-
holtz) or the degree of adaptation or fatigue in the receptor
system (Delboeuf), however, c is a free parameter.
Two other intermediate functions (see Krueger 1989a, sect. 1)
are the Naka-Rushton function (Lipetz 1971), which Murray
(see his Figure 3) also fitted to Delboeuf's data, and Norwich's
(1987) information-theoretic entropy reduction measure. No
doubt there are several other intermediate functions lurking
about as well. Thus, I don't see anything particularly special
about Delboeuf's function, and Murray concedes "that curve-
fitting alone is unlikely to decide between psychophysical func-
tions as applied to real data in certain cases" (Note 3).
Murray nevertheless seems caught up in what I regard
(Krueger 1989a, sect. 9.2) as the vain search for "the" psycho-
physical function. The sad fact is that we do not yet know (and
may never know) which trends (linear, quadratic, cubic, etc.) of
the underlying function are truly significant and ought to be
captured by the psychophysical equation. In the interim, scales
ought to be dealt with on a holistic basis, rather than in terms of
the parameters of a particular function. If Fechner had been less
committed to the logarithmic function, and had dealt simply
with summated jnds (as he himself realized was possible; see
target article, sect. 1.2), he would presumably have accepted
the more expansive scale that the near miss to Weber's Law so
strongly implies (Krueger 1989a).
3. Nature of variability. Although variability is a pillar of "inner
psychophysics' according to Murray's Figure; 1, it is not clear
from Murray's History whether the founders of psychophysics
had a deep understanding of the nature and meaning of vari-
ability. Variability is reducible insofar as it represents the effect
of variation on uncontrolled factors (see Murray's citation of
Boring in sect. 1.4.1). Thus, Stevens et al. (1941) obtained a
distinct type of variability - which they associated with the
operation of discrete neural quanta - after they removed or held
constant many such factors. Even their remaining variability
might have been eliminated if further controls had been used,
such as locking stimulus presentation to particular brain states.
A key question is whether any portion of the variability in
psychophysics, like the uncertainty associated with quantum
mechanics, is truly irreducible, that is, based on probabilistic
(static) or stochastic (dynamic) "noise" factors rather than deter-
150 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
ministic ones. Chaos theory (Gregson 1991; Krueger 1991)
shows that even simple processes may produce complex behav-
ior and that deterministic factors may produce unpredictable
results, owing to extreme sensitivity to initial conditions. Thus,
not all that appears to be random may be truly random
(nondeterministic).
The Weber fraction, k, in A/ = kl (see Murray's Equation 1),
is larger in some modalities, indicating increased variability in
observers' behavior. The reciprocal of this factor, Ilk, is a
component of the measure constant, C, for Fechner's log-
arithmic function (see Murray's Equations 2-4). Thus, for Fech-
ner, the subjective magnitude, S, ought to rise faster with
objective magnitude, /, when discriminability on the dimension
is poor (i.e., large k and thus small C). However, as far as I know,
Fechner never made anything out of this fact; Murray (sect. 1.1)
said that nowhere in his three books "does Fechner give a graph
or a table showing sensation magnitudes as a logarithmic func-
tion of stimulus intensity."
Fechner gave more attention to the absolute threshold, /,
which anchors his scale (see Murray's Equation 4), than to the
difference threshold or just noticeable difference (jnd), whose
size determines k and thus C. Thus, it would appear that
Fechner was not very interested in certain aspects or forms of
variability, perhaps regarding them as due to second-order
nuisance factors. This indifference may be well founded. Ac-
cording to the proportional jnd theory (Houtsma et al. 1980; Lim
et al. 1977; Riesz 1933), the subjective size of the jnd depends on
m, the total number ofjnds in a particular modality, and thus the
\lk component of the measure constant, C, ought to be replaced
by l/mk (see Krueger 1989a, sect. 8). If, in addition, k and mare
reciprocally related or constrained, that is, ink = In (dynamic
range) (Norwich 1987), then "resolving power per se may have
nothing to do with the psychophysical power function" (Krueger
1989a, sect. 9.2, p. 264).
Fechner anchors his scale at what is really a statistical fiction;
his anchor, the absolute threshold, / 0 , is operationally defined as
the level of/ producing 50% detection, and thus it is more an
experimenter's "criterion" than an observer's "threshold." In
fact, owing to variability in the observer, positive sensations
may be experienced well below the absolute threshold, contrary
to Fechner's Law, which holds that when / = /, S must equal
zero. Fechner faulted the power law because in its unmodified
form (target article, sect. 1.3.3), "its assertion that if I were
nonzero, the subject would necessarily have a sensation S of
some strength." Given the intrinsic noise of the receptive
system (sect. 2.1), as evident in the "natural light of the retina"
(Helmholtz), however, there may generally be a positive sensa-
tion S even when / is zero, not to speak of when / is above zero
(e.g., / = U
4. Physical scale. Fechner's lack of concern regarding the size
of the measure constant, C, is shown again in his indifference
regarding the choice between energy and velocity for the
physical scale. He stated that the measure constant is simply
doubled in the case of velocity (see Fechner quote in Murray's
Note 1). Most psychophysicists seem to share Fechner's indif-
ference about the choice of physical scale, judging by the scant
treatment accorded this issue in Murray's History. It is a serious
issue nevertheless (Krueger 1991, Note 2; Myers 1982; Weiss
1981; 1989), and may be just as important as the issue of rating
method because the exponent of the power function can be
doubled or halved depending on the choice of physical scale
(amplitude, power) as well as on the choice of rating method
(magnitude estimation, category rating).

The antecedents of signal detection theory
Donald Laming
Department of Experimental Psychology, University of Cambridge,
Cambridge CB2 3EB, England
Electronic mail: drjl@phx.cam.ac.uk
Murray's Figure 1 links some of the more important studies in
psychophysics according to their "commonalities of interest"
rather than any "direct historical interaction. " A casual reader
might expect each column to be addressing a common problem,
each study building on the work of its predecessor; he might
take the historical span of Murray's diagram to indicate the
profundity of Fechner's work, shaping the development of
psychophysics for 130 years. But the column headed "Vari-
ability," to take one example, simply lists studies that have
hypothesised about the internal variability of sensory processes,
irrespective of any relation between them or of where their
respective ideas originated from.
1. Tracing antecedents. This commentary shows by example
that tracing the sources of the essential ideas in each article leads
to quite a different diagram (see my Figure 1). By "essential," I
mean those ideas that enable theory tofitdata - which make the
sums add up - without which there would be no theory. Now, in
psychology at least, the presentation of theory is usually accom-
panied by images or examples that, although not themselves
part of the theory, provide a vehicle for the reader's thoughts - a
kind of wrapping paper for the psychological goods. Signal
detection theory will provide an instructive example. I am
enquiring here about the origin of the goods, not of the wrap-
ping paper, no matter how fashionable that might be.
1.1. Three principles. I set out three principles for tracing the
origin of ideas:
(i) The trace goes backwards in time because references to
other publications are necessarily retrospective. I shall begin
from the principal publications in signal detection theory, not
from the sources of the ideas.
(ii) Essential ideas are most easily identified in the earliest
publications. Later publications (e.g., Swets et al. 1961) are
more elaborately wrapped, with references to other analogous
ideas included to make the theory easier to assimilate. I shall
begin my inquiry from Tanner and Swets (1954).
(iii) Attention must be focussed on those ideas actually used in
calculation. Quantitative theories are therefore easier to trace.
2. Signal detection theory. Tanner and Swets (1954) list only
two references. One (Blackwell et al. 1954) concerns only the
apparatus used in their experiments. The other, a technical
report subsequently published as part of Peterson et al. (1954),
"The theory of signal detectability," is the immediate source of
all the theoretical ideas.
2.1. The theory of signal detectability. Peterson et al. (1954)
addressed the specifically military problem of optimum detec-
tion of radar signals. They list 25 references, of which I judge
three to be essential (to theory rather than practice).
A radar signal is picked up as a continuous waveform over
some interval of time, confounded with background noise.
Some mathematical theory is needed to support a discussion of
the probability attaching to one sample waveform rather than
another. That theory comes from two sources: the analysis of
random signals and noise by Rice (1944/45) and the sampling
theorem of Shannon (1949).
Once probabilities have been defined for sample waveforms,
an optimal basis is required for deciding whether or not a true
signal is incorporated in the waveform. That optimal basis is
provided by the Neyman-Pearson Lemma (Neyman & Pearson
1933).
2.2. The Neyman-Pearson Lemma. This lemma states, as a
general principle, that any "most efficient statistical test proce-
dure" is a simple dichotomy of likelihood ratio (in favour of one
state of nature and against another). Neyman and Pearson (1933)
Commentary/Murray: History of psychophysics
appear to have had no essential sources other than their own
earlier papers (e.g., Neyman & Pearson 1928), except that the
idea of a ratio of likelihoods came from Fisher (1922, p. 326).
These antecedents of signal detection theory are summarised
in Figure 1. I emphasise that they are all extraneous to experi-
mental psychology. In particular, although Thurstone (1927a;
1927b) is referenced by Swets et al. (1961), that reference
signifies no more than a superficial similarity. It was not a source
of ideas. Thurstone did not add a yes/no decision rule to his
model (target article, sect. 1.4.1) because that particular idea
did not arrive until six years later (Neyman & Pearson 1933). He
did not develop his "law of comparative judgment" into a signal
detection theory because his eyes were set on scaling attitudes
and other social entities (Thurstone 1929; Thurstone & Chave
1929).
3. The problem of "variability" -1. Murray's approach to signal
detection theory focusses on its wrapping and entirely passes
over its essential contribution to psychophysics. To make this
point Figure 2 reproduces data from Swets et al. (1961, Figure
13, Observer 1). The empirical reality of a signal detection
experiment is a set of data points (here, five of them) exhibiting
an essential relation between a, the probability of wrongly
reporting a signal when noise alone has been presented (corre-
sponding to the significance level of a statistical test), and B, the
probability of reporting a signal when one is present (the power
of the test). In measuring detectability there are two proba-
bilities to be estimated. If one of them (a) is left uncontrolled, as
in most experiments prior to 1954, the data are confounded.
Signal detection theory was a revolutionary idea. It showed
psychophysicists, for the first time, how experiments on detec-
tion and discrimination should be designed and analysed, and
why. There was, indeed, a profound problem of "variability" (a
uncontrolled), quite distinct from the problem Murray (sect.
1.4.1) has in mind.
3.7. The problem of "variability" 2. Murray (sect. 1.5.1) exam-
ines the possible relation of the normal distributions of signal
detection theory to underlying brain activity. That falls into the
category of wrapping paper for this reason:
Any theoretical operating characteristic can be succinctly
written in parametric form,
P("Yes" | Noise) = o(i)
P("Yes" | Signal + Noise) = |}(x). (1)
In terms of a conventional signal detection model, x is the
decision variable and l-ot(x) and l-B(x) are the cumulative
distribution functions. The point to be emphasised is that
Equation 1 continues to describe the same operating charac-
Likelihood ratio
(Fisher 1922)
Random signals and noise
(Rice 1944/45)
Sampling theorem
{Shannon 1949)
Neyman-Pearson Lemma
[Neyman Pearson 1933)
Theory of signal detectability
(Peterson, Birdsall Fox 2954)
Signal detection theory
{Tanner & Swets 1954)
Sensory analysis
(Laming 1986)
Figure 1 (Laming). A chart tracing the origins of explanatory
ideas in signal detection theory.
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 151
Commentary/Murray: History of psychophysics
1.0
_ *-
0.8-
.*
in_ 0.6-
"w
t, 0.4-
0.2-
Observer 1
0.0
0.0 0.2 0.4 0.6 0.8 1.0
P("Yes"|N)
Figure 2 (Laming). Signal detection data from Swets et al.
(1961; Figure 13, Obs. 1) fitted with a noncentral x 2 operating
characteristic (Laming 1986, p. 258).
teristic, the same empirical reality, under any monotone trans-
form of x. There are many different pairs of distribution func-
tions, l-a(ac) and l-P(x), concordant with the same operating
characteristic. The idea that those distributions are normal is
pure assumption. The experimental data do not tell us what
shape either of them has; the data speak only to a relation
between the two. The interpretation of specifically normal
distributions in terms of brain activity addresses an artificial
question.
The practical choice of distributions is wider than Equation 1
allows because empirical data are subject to sampling fluctua-
tions and are compatible (to within the limits of experimental
error) with a variety of different operating characteristics. To
emphasise this point, the broken curve in Figure 2 is calculated
from a noncentral x2 model (Laming 1986, pp. 95 & 254-60). It
fits well (x2 = 1.932, 3 d.f).
3.2. The legacy of "signal detectability." Normal distributions physics to adhere rigidly to the framework set out by Fechner
have become attached to signal detection theory because they
constitute the simplest of the specific cases ("Signal known
exactly") worked out by Peterson et al. (1954) and thefirstmodel
to be explored by Tanner and Swets (1954). Others of Peterson
et al.'s specific cases are reproduced in Green and Swets (1966,
Ch. 6) and provided hypotheses for many of the early experi-
ments. But the "theory of signal detectability" calculated the
possible performances when the statistics of the signal and noise
were known. Signal detection theory faces the complementary
problem: The performance of a human observer can be recorded
experimentally; what might be the statistical properties of the
underlying signals?
3.3. Sensory analysis. In my Sensory analysis (Laming 1986; In his target article, Murray seems to have missed one interest-
see also multiple book review BBS, 11[3] 1988) I addressed the
questions of why signal detection operating characteristics have
the shapes they have; why those shapes differ, comparing a
discrimination between two separate stimuli with the detection
of an increment added to a background; and how those shapes
depend on physical parameters of the stimuli. The mathematics
of random signals and noise, one of the essential constituents of
the theory of signal detectability, but of no more than peripheral electroencephalogram. Every subject has his own individual
concern to signal detection theory, again plays an essential role dominant frequency of alpha rhythm in the 8-13 cycles per
because the sensory processes on which human discriminations
are based evolve in continuous time. The manner in which
probabilities can be assigned to samples from such processes is
an essential part of the theory.
Sensory analysis stands in a direct line of descentfromsignal
detection theory and its antecedents. Murray (Fig. 1) puts it in
quite a different part of his diagram.
4. Fechner's heritage. Murray's tracing over an interval of 130 perceptual delay depends inversely on stimulus intensity
152 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
years of themes first discussed by Fechner might be taken as
evidence either of Fechner's genius in identifying fundamental
questions or of a lack of competence on the part of his successors
in answering them. I think both suggestions are wrong. Instead,
Fechner has directed attention to some inappropriate questions
(to which answers are never found); and his successors have
become fixated on those particular questions, unable to turn
away, even when experimental evidence so dictates. I give two
examples.
Swets et al. (1961) were writing against an established con-
sensus based on the notion of afixedthreshold. They presented
several experiments, most notably their second guess experi-
ment (pp. 321-24), showing that idea to be untenable. The
following years saw a number of alternative signal detection
models incorporating that untenable idea (Atkinson 1963; Black-
well 1963; Krantz 1969; Luce 1963; Norman 1964), addressing
(of course), not the data which showed a fixed threshold to be
untenable, but other data, too imprecise to be discriminative.
More recently Quick (1974) has proposed this function,
p= l-exp{-(sx?}. (2)
to describe the probability that thefixedthreshold of a detector
of sensitivity s is exceeded by a stimulus of magnitude x. The
desirable property of Equation 2 is that the corresponding
equation for the probabilistic combination ("probability summa-
tion") of independent detectors of possibly different sensitivities
but the same exponent k has the same exponential form, which
is, moreover, easy to calculate. It has been much used in the
modelling of psychometric functions for the detection of grating
stimuli (e.g., Graham 1977; Legge 1978; Robson & Graham
1981; Watson 1979); and Foley and Legge (1981) have analysed
their data with respect to both Equation 2 and a normal proba-
bility integral with respect to some power (k) of the stimulus
magnitude (such as one might derive from signal detection
theory; see Laming 1986, Ch. 4). The comparison is not
illuminating.
What these examples illustrate is the parochial practice of
scientists who focus on their own narrow field of research only
and ignore discoveries outside. It enables research in psycho-
and it limits progress to the purely lateral.
Derivation of Stevens's exponent
from neurophysiological data
Artour N. Lebedev
Institute of Psychology, Russian Academy of Sciences, 129366 Moscow,
Russia
Electronic mail: farib@ipras.msk.su
ing concept in psychophysics, the concept of time quanta
(Geissler 1990). These are closely related to the periodicity of
brain processes, which constitute the basis of mind in accor-
dance with the ideas of David Hartley (1967) and Gustav
Fechner (1851).
One typical time quantum is about t = 0.1 sec. This is the
mean period of the most regular alpha rhythm in the human
second range. A small step-like difference between neighbour-
ing peaks in alpha waves is equal to the relative refractoriness
after each neural impulse. This is another time quantum, r =
0.01 sec.
In our previous work (Lebedev 1990; Lebedev & Myshkin
1989) we computed maximum perceptual delay Tmax = t(t/r)
and minimum perceptual delay Tmin = r/2. This changeable

(Pieron 1914) with an exponent the absolute value of which is
approximately equal to the exponent in Stevens's Law. This
relates to brightness and loudness (Bonne 1989).
Equal perceptual delays correspond to equal sensations. One
can therefore think that the range of subjective sensation
strength (S) evoked by different intensities (/) is equal to the
range of changeable perceptual delays:
2P _ Tmax _ Smax _ / /max \"
r2 Tmin Smin V /min / (1)
where n is Stevens's exponent and max and min are indices of
maximum and minimum values.
As a sequence,
n = (2*log(t/r) + Iog2)/log(/max/7min). (2)
This formula reflects the known inverse relationship between
stimulus range and Stevens's exponent (Sarris 1967; Teghtsoo-
nian 1971). The value of the exponent is determined by typical
neurophysiological parameters. The maximum range of per-
ceived intensities is limited by the maximum diversity of all
possible neural codes
/max __ tt_ _ i
/min \r (3)
where the base is an alphabet span of groups of neural impulses
and the exponent is the number of such overlapping groups
appearing sequentially during a single cycle of neural activity.
Jarvilehto (1982), working with single tactile receptors, has
shown that the magnitude estimate actually grows in linear
relation to the number of impulses within a single group. Jointly
solving equations (2) and (3) yields n Ss 0.3 which accords with
the experimental data of many authors.
The right half of formula (3) simultaneously limits the diver-
sity of all possible units of perception and memory. It has been
shown that short-term memory span, visual and memory search
speed, and latency of simple and choice reaction time depend on
the above-mentioned neurophysiological parameters (Lebedev
1990; Lebedev & Myshkin 1989). Hence, psychophysical regu-
larities closely interrelate with cognitive ones through common
neurophysiological information-processing parameters. The
core problem of "sensation strength" underlined by Murray will
be exhaustively solved if one keeps in mind the neuro-
physiological regularities of cyclic brain processes which indeed
belong to inner psychophysics.
Fechner's theory of mental measurement
Stephen Link
Department of Psychology, McMaster University, Hamilton, Ontario,
Canada L8S 4K1
Electronic mall: link@mcmaster.ca
The appearance of Murray's broad perspective is yet another
sign of a renewed interest in psychophysical inquiry. No doubt
there are as many perspectives of this reemerging elephant as
there are viewers, for example, Scheerer's (1992) view of Fech-
ner's inner psychophysics. From my vantage point there is no
more breathtaking sight than Fechner's creation of scien-
tific psychology. The result was the foundation for much of
nineteenth- and twentieth-century psychological research and,
as statisticians acknowledge (David 1969; Neyman 1976), a
foundation for statistical hypothesis testing and descriptive
statistics. Because this foundation for psychological research is
largely forgotten, I present it as an addendum to Murray's
perspective.
Fechner's idea was powerful and simple. Suppose one claims
to possess the mental ability to feel weight. To ascertain the
Commentary/Murray: History of psychophysics
extent, if any, of this ability one is asked to judge which of two
objects weighs more. Fechner assumed that each object, when
lifted, produced sensations of heaviness that were mentally
combined. In this way the average of two heavinesses became a
standard against which an individual heaviness was judged
(Fechner 1860/1966, pp. 85-89). In making a judgment of
which of two stimuli weighed more, the heaviness produced by
one object was compared against the average. When an object's
heaviness was perceived to be greater than the average, that
object was judged to be the heavier of the two. When less, the
object was judged to be the lighter.
Fechner, once professor of physics at Leipzig, supposed that
the sensory system evoking the sensation of heaviness obeys the
same law of error that affects any physical measuring device.
According to Gauss's (1809; 1821/1880) Law of Error, there
exists a true value of the object to be measured and an obtained
value that is the sum of the true value and a randomly deter-
mined amount of error. Thus, the sensation of heaviness evoked
by a constant physical stimulus varies depending upon the
amount of error added to the true heaviness. Whenever the
lighter of the two weights was judged to be the heavier, or vice
versa, an incorrect judgment occurred that was attributed to the
role of error.
Measuring the size of this invisible sensory error in units of
the visible physical stimuli was one of the most outstanding
scientific achievements in the nineteenth century. Fechner
argued that the difference between two stimuli, say weights, is
measured in grams. These two stimuli also produce a difference
between sensations measured in units of sensory error. Thus,
the number of sensory error units separating the two sensations
equals the number of grams separating the two weights. Divid-
ing the number of sensory error units into the number of grams
determines the unit of sensory error in grams. Thus the differ-
ence between the two physical stimuli is measured with respect
to sensory error units, that is, with respect to the resolving
power of the mind. In this way Fechner and his psychophysics
exposed to scientific scrutiny differences between sensations
that were previously hidden from public view and observed only
by personal awareness.
The theory of this measurement, based upon derivations
corroborated by the well-known topologist A. F. Mobius (b.
1790 - d. 1868), will appear obvious to students of classical
statistics. It is illustrated in Figure 1, where the abscissa repre-
sents heaviness. On the abscissa are the true heaviness values, a
FECHNERIAN THEORY OF DISCRIMINATION
75% CORRECT DISCRIMINATIONS
Distribution of
Heaviness for S A
RESPOND "HEAVIER"
^^S a Errors \ ^
RESPOND "LIGHTER"
/
r\ \ \ Distribution of
>. Heaviness for S_
b
^ ^ Errors
- 4 - 3 - 2 - 1 0 1 2 3 4 5 6
TRUE HEAVINESS + GAUSSIAN ERROR
Figure 1 (Link). Two Gaussian probability density functions
represent the distribution of possible heaviness values pro-
duced by lifting stimuli SA and SB. Fechner assumed that a
person required to judge which stimulus was heavier created a
threshold for judgment halfway between the means of the two
distributions. (Substantial parts of this comment and Fig. 1 are
reprinted from Link 1992, with permission).
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 153
Commentary/Murray: History of psychophysics
and b, for two weights corresponding to stimuli SA and SB. The
average of these two values (a + b)l2 Fechner referred to as die
schwelle, or the threshold. Fechner postulated that when the
lighter of the two weights SA was judged to be the heavier of the
two, it was because the true heaviness of the lighter weight had
added to it enough error to make its apparent heaviness greater
than the average heaviness of the two weights. Because the
lighter weight was perceived to be heavier than the average, it
was judged, incorrectly, to be the heavier.
By hypothesis, when stimulus SA was lifted the sensory
system added a Gaussian error having mean value zero to the
true heaviness, a. When an error occurred, the heaviness of SA
necessarily exceeded the rerjrent located at (a + b)l2. Because
(a + b)l2 also equals a + (o - a)/2, Fechner reasoned that the
amount of error was larger than (b a)/2. Therefore, the
probability of an error ofjudgment must equal the probability of
errors larger than (b a)l2. The shaded area under the Gaussian
probability distribution to the right of the midpoint, (a + b)/2,
represents the probability that SA has sufficient error added to
its true heaviness to exceed the average heaviness of the two
weights and produce an error of judgment.
When the larger weight, SB, was judged to be the lighter of
the two, the heaviness was judged to be smaller than (o + b)l2,
which equals b (b o)/2. The total heaviness was less than b
by an amount at least as big as (b d)/2. Thus the error was of
the same magnitude for SB as for SA. Due to the symmetry of the
Gaussian distribution and the use of the midpoint as a threshold
for judgment, the probability of an error in judging SB equals
the probability of an error in judging SA.
To estimate the probability of an error, Fechner determined
the relative frequency of an error by computing the ratio of the
observed number of such errors divided by the total number of
judgments where such errors could occur. Therefore, an error
larger than (b a)l2 occurred with a probability that was
determined through mental judgments - self-reports about the
status of a sensation. Fechner suggested that by repeated
stimulus presentations enough judgments could be gathered to
provide a statistically reliable estimate of this unknown proba-
bility. For this reason the probability of an error in judgment
became an important element of the theory of measurement.
For Fechner, estimating the probability of an error of judg-
ment was only a step along the path toward mental measure-
ment. His major concern was to measure the sensitivity of the
sensory system in much the same way as determining the
sensitivity for any physical measuring device. In the mid-
nineteenth century the statistical quantity called the probable
error served this purpose. The probable error equals the dis-
tance from the mean of a Gaussian probability distribution to a
point that yields an area under the Gaussian distribution equal
to 0.25. Plus or minus one probable error from the mean defines
the interquartile range containing 50% of the Gaussian proba-
bility distribution. Beyond one probable error above the mean
is 25% of the distribution which corresponds to large positive
deviations from the mean. Another 25% of the distribution
consists of large negative deviations that exceed one probable
error below the mean.
Suppose two weights of 300 g and 318 g are correctly discrimi-
nated with probability 0.75. The mean heavinesses of these
stimuli are shown as points a and b on the abscissa of Figure 1. In
terms of decision making, the average heaviness is a decision
criterion that lies midway between the heavinesses correspond-
ing to the two stimuli. The average heaviness corresponds to an
interpolated hypothetical stimulus of 309 g. The probable error
equals 9 g because 25% of the probability distribution of heav-
inesses exists between 300 g and 309 g.
To measure the sensitivity of the sensory system, Fechner
used the reciprocal of the probable error. As the probable error
increased, and larger errors became more likely to be added to
the true heaviness, this reciprocal decreased, indicating a re-
duction in sensitivity. For a small probable error, few large
154 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
deviations from the true heaviness occurred and the measure of
sensitivity increased.
The measurement of the probable error, based on Herbart's
and Weber's idea of a threshold, allowed psychological scientists
to determine whether one judge could discriminate a smaller
difference between two stimuli than another, or whether, for
the same judge at different times, the size of the physical
difference needed for 75% correct discrimination varied. In this
way Fechner showed how to investigate individual differences
and also how to examine variations within an individual to such
causes as fatigue, time of day, and order of stimulus presen-
tation.
The major point of Fechner's psychophysics is that the wholly
invisible error of the sensory measuring device was made
measurable through an extension of Gauss's Theory of Error.
The theory was later recreated by engineers who were entirely
unaware of Fechner's contribution to the founding of psychol-
ogy. This new rendition is a form of signal detection theory
known as ideal observer theory.
Somehow this quite stunning idea became lost to psychology.
Many students trained by Wundt in Leipzig took up residence
in the United States. Yet their textbooks did not present
Fechner's theory. Perhaps the loss of Fechner's theory is a
consequence of the loss of many young European students in
two World Wars, compounded by the zealous grip on experi-
mental psychology of the antiquantitative, behavioristic forces
of the twentieth century. The antiquantitative bias is revealed in
a telling comment found in the introduction to Experimental
psychology: A manual of laboratory practice by E. B. Titchener
(1901), scion of Wundt's line in American experimental psychol-
ogy: "Knowledge of elementary mathematics is a part of a man's
general scientific outfit: but one may work a lifetime, and with
success, in psychology, without needing the knowledge" (p.
xxiii).
Fechner's great insight did not entirely disappear, but the
band of psychological scientists dedicated to the measurement
of psychological phenomena grew smaller as the statisti-
cal/mathematical models of additive effects came to dominate
many psychologists' methodology. The story of the theorists and
experimentalists who continued beyond Fechner's pioneering
work is told in Part I of Link (1992, "Psychophysical origins of
difference and similarity"). Part II presents the theory that
generates feeling from sensation and resolves many previous
psychophysical paradoxes by showing how Weber's constant,
Fechner's Law, and Stevens's exponents are derived and
related.
A parallel view of the history
of psychophysics
Gregory R. Lockhead
Department of Experimental Psychology, Duke University, Durham, NC
27706
Electronic mail: dgreg@dukemvs.bitnet
Murray has outlined a history of psychophysics in terms of
Fechner's "outer" and "inner" psychophysics. To make that
history more complete, I would add a column to his Figure 1 and
entitle it "Contextual Psychophysics," to reflect Egon
Brunswik's history of psychophysics (1956, Figure 7, p. 32).
Like Murray, Brunswik began his history of psychophysics
with the classical approaches of the early and middle nineteenth
century. Brunswik's history described relations between exper-
imental design and data interpretation. As an example of the
beginnings of psychophysics, Brunswik reviewed studies of the
Galton Bar that provided threshold measures in terms of differ-
ence limens and PSEs, examined generalizations of the Weber-
Fechner law, and introduced new statistical measures.
Brunswik called this "structuralism." It was alleged then by

investigators that individual stimulus variables could be isolated
and measured.
It gradually became clear that factors other than the variable
of interest determine the perception of that variable. To make
these demonstrations, that attribute was systematically isolated
from neighboring variables. As examples, Brunswik reviewed
studies of the Miiller-Lyer figure which showed that the appar-
ent length of a line depends on what lines surround it. Such
research was an important part of the beginnings of Gestalt
psychology.
It further came to be understood that such multidimensional
determiners of perception were related to thing-constancies.
This was shown in studies that systematically isolated distal cues
from other mediating cues. The results are that a distal stimulus
appears constant even though the proximal stimulus varies
widely. This fact requires relaxing the idea that the intensity of
an attribute at the receptors determines its perception and that
the way to measure appearances is to hold constant everything
in the situation except that attribute.
Brunswik next extended his analysis to emotions and attitudes
and reported studies showing how these are also integral to
judgment. His consistent emphasis was on the biological value
of the situation rather than on the intensity of an attribute at the
receptors.
This does not mean that his approach could not determine
thresholds or that he considered such measures unimportant.
Neither would be true. But it does mean that Fechner-Iike laws
cannot be correct beyond being points in a complex function
where performance depends on many factors which sometimes
interact in complex ways (Lockhead 1992).
Not everyone was persuaded by Brunswik and separate
literatures resulted. As examples, Brunswik did not reference
Delboeuf or Plateau, who are important to Murray's review, and
Murray did not reference Brunswik or most of the other re-
searchers who are important to Brunswik's review.
In addition to different methods, different dependent vari-
ables are also sometimes used. In a sample "outer psycho-
physics" study, the brightness of a disc is measured when its
intensity is varied and the surround is held constant. The result
is that disc brightness increases with 'intensity. In a sample
"inner psychophysics' study, as reviewed by Murray, the over-
all illumination of the entire display is varied and the perceived
contrast between the disc and surround is evaluated. Brightness
is not reported here, and perceived contrast is not noted in disc
brightness studies.
The constancy literature shows that disc brightness changes
little if at all when the overall illumination changes a lot, even
though disc intensity changes a lot (Brunswik 1956). These
conflicting conclusions - brightness increases with intensity
(when disc intensity increases) and brightness is independent of
intensity (when disc and surround intensity both increase) -
argue against the classical assumption that individual variables
can be isolated and measured. Something beyond the models
reviewed by Murray is required for completion.
Because of such considerations, Brunswik concluded that our
understanding of perception must move away from the classical
considerations, which examined effects of the proximal stim-
ulus, to considerations of constancy, which examined relations
between perception and the distal stimulus. This was a move
from a structuralism that emphasizes effects of stimulus inten-
sity at receptors to a functionalism that emphasizes the biolog-
ical importance of perception.
In relating method to theory, Brunswik contrasted proce-
dures used by classical (mid-nineteenth century) researchers
who were interested in direct intensity effects with procedures
used by then modern (1890s) experimentalists who were inter-
ested in appearances. The classical methods held everything
constant except the attribute of interest. The "moderns,' at-
tempting to mimic the ordinary world, held essentially nothing
constant.
Commentary/Murray: History of psychophysics
Brunswik suggested using the modern procedure plus the
"omission technique' (Brunswik, p. 25) of gradually eliminating
cues as information accumulates. Because we do not know a
priori what determines judgment, this method allows poten-
tially essential cues to be preserved and thus possibly to be
discovered.
Classical experimentalists bypass this process because they
already "know" that attribute intensity determines perception.
When the assumption is correct, this is the more efficient
method. When it is wrong, one might expect irregular data,
so the erroneous idea would quickly be rejected. Unfortunately,
the data are sometimes confounded with what actually causes
the performance so that face validity gives encouraging results
with no motive to look further.
Consider typical brightness studies. It is not possible to vary
only the intensity of a spot. Then the difference between the
spot and its surround also changes. Furthermore, brightness is
determined by this difference and by the overall illumination,
not by spot intensity (Lockhead 1992). Using only the classic
approach and assumptions, this remains unknown because
there is no apparent need to seek a causal factor beyond spot
intensity.
Because Brunswik recognized this as a potentially general
difficulty, his history of psychophysics applauds what he saw as
progress in moving from the classic view of independent attri-
butes to a Gestalt psychology and then to a functional approach
where constancy is the important fact and the study of percep-
tion calls for an overt, distal, and multidimensional psycho-
physics based on the idea that what is most important is "the
relationship between ecological validity and utilization"
(p. 141).
Let's not promulgate either Fechner's
erroneous algorithm or his
unidimensional approach
R. Duncan Luce
Department of Cognitive Sciences and Institute for Mathematical
Behavioral Sciences, University of California, Irvine, CA 92717
Electronic mail: rdluce@uci.bitnet
Murray several times cites Luce and Edwards's (1958) critique
of Fechner's derivation of Fechner's Law from Weber's Law
(sects. 1.1, 1.2, and Note 2). Unfortunately, he seems not to
have absorbed our message and he makes the same mathemati-
cal error where it matters some. Fechner correctly stated
Equation 2, AS(I)/A7 = Cll, as the functional equation embody-
ing both his hypothesis of equal subjective jnds and the assump-
tion of Weber's Law. To solve for S(/), he proposed the algo-
rithm - calling it a "mathematical auxiliary principle" - of
replacing AS(/)/AZ by the differential dS(l)/dl. Thus, Equation 2
is transformed into a (well-understood) differential equation
whose solution is Equation 4. Our point was that while this
algorithm happens on a correct solution when either Weber's
Law or the generalization implicit in Equation 5 holds, it fails to
arrive at the correct solution for any other Weber function.
Falmagne (1985) provides a good discussion of a correct
algorithm.
This may seem a bit of esoterica, but the issue reappears in a
more significant way in section 1.3.3 where Murray notes that
Fechner later considered a reformulation (see Equation 10) of
the initial problem. To "solve" Equation 10 using Fechner's
algorithm, rewrite it as AS(/)/A7 = CS(1)/1. Replace the ratio of
differences by a differential to get dS(l)/dI = CS(I)/I. This is
equivalent to d[\og S(I)]/dI = Cd(log l)ldl. Integrating and
taking exponentials yields Equation 11. Murray appears to
accept this not only as historically accurate, but also as mathe-
matically correct. About the latter he is wrong. One disconcert-
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 155
Commentary/Murray: History of psychophysics
ing feature of the "solution" is that it is independent of the form
of the Weber function. Another worse feature is that Equation
11 simply does not satisfy Equation 10 except when C = 1. To
show this, let/' = / + A/, andsoAS(/) = S(J') - S(I). Assuming S
is given by Equation 11,
=m _1= [AI
1
I* _
s(i) sy)
L/ J theory. Summaries of various of these models can be found in
which agrees with Equation 10 when and only when C = 1.
Considering that 34 years have passed since it was noted that
substituting a differential for a ratio of differences in general
leads to an incorrect solution, it is sad that this not very subtle
mathematical issue continues to mislead.
A conceptual issue permeates Murray's target article. Is it
plausible to expect successful measurement of an attribute when
just one independent variable is manipulated? To my knowl-
edge, the only case where this approach has succeeded is for the
extensive measures of physics, of which mass, length, charge,
and time are prime examples. Such dimensions include a
concatenation operation of combining two entities that each
exhibit the attribute in question to form a third entity that also
exhibits the attribute. Classically, such operations were repre-
sented numerically as addition, although that choice is highly
conventional - multiplication is equally good, as are a contin-
uum of other (associative and commutative) mathematical oper-
ations. No other purely one-dimensional example of fundamen-
tal measurement is known, and that was the reason N. R.
Campbell (1920/1957) and the Ferguson (1940) committee,
largely a creature of Campbell's concerns and on which Camp-
bell played a major role, concluded that psychology was incapa-
ble of fundamental measurement: It has no empirical concatena-
tion operations of its own.
If one accepts that measurement is a one-dimensional matter,
the committee was right. x For a careful contemporary treatment
of the one-dimensional approach and its ambiguities, see
Narens (submitted). Falmagne (personal communication) points
out that if one is willing to deal with choice probabilities rather
than orderings, it is possible in principle to construct a binary
operation over the probability space. The difficulty is that
because most choice probabilities are 0 or 1, one is forced to
piece together the global scale from highly local data.
What the Ferguson committee failed to acknowledge, and
many psychophysical sealers seem to continue to ignore, is that
something very like one-dimensional measurement becomes
feasible when two or more independent variables affect the
same attribute. One can use the resulting trade-off between the
independent variables as a source of measurement of the attri-
bute and how the factors combine. Indeed, trade-offs typically
induce mathematical concatenation operations on the compo-
nents. This was completely familiar in physical measurement of
such quantities as momentum, energy, density, and so on, but it
had not been axiomatized in a fashion analogous to the turn-of-
the-century axiomatizations of extensive measurement. The
lacuna was corrected in the early 1960s (see Chapter 6 of Krantz
et al. 1971 and Chapter 19 of Luce et al. 1990 for historical
details).
Thoroughgoing behavioral examples of the trade-off approach
are Luce's (1977) axiomatization of power and logarithmic func-
tions when the data are orderings of stimulus pairs, such as
loudness to binaural tones; the functional measurement proce-
dures advocated and applied to psychophysical as well as other
psychological problems by Anderson (1981; 1982); and the
entire complex literature on axiomatizations of preferences
and/or judgments of uncertain alternatives. In each case, one
uses the trade-off between variables to establish simultaneously
the measures involved and the law relating them.
If axiomatics are not to one's taste, another approach is to
increase the dependent variables from just choice and/or judg-
ments about psychophysical stimuli to include the time taken to
156 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
respond. Here the models are much more process oriented and
have little in common with the measurement-theoretic models.
The measures of sensation are parameters of the model and are
estimated from behavioral data. Again, trade-offs are the name
of the game, in these cases often between two or more sensory
variables such as signal duration and signal intensity but also
between a sensory variable and some sort of motivational crite-
rion. The models simultaneously develop measurement and
Luce (1986) and Townsend and Ashby (1983).
My only point is that one is probably misguided to continue to
fuss at the one-dimensional measurement case; unless a relevant
concatenation operation or some other rich internal structure
can be found, the situation is simply too underdetermined to be
of much theoretical interest.
NOTE
1. At the time, the only cases that were understood involved opera-
tions with additive representations. Later work (see Chs. 19 and 20 of
Luce et al., 1990, for a summary) yields a whole family of inherently
nonadditive operations. Homogeneity, described in Luce and Narens
(1987), underlies this result; this paper, which is purely expository,
gives references to the original contributions.
Quantifying, valuing, choosing
Lawrence E. Marks
John S. Pierce Laboratory and Yale University, New Haven, CT 06519
Electronic mail: marks@yalevm.bitnet
In his thoughtful historical review, Murray recounts several
issues that were raised to prominence by Fechner and others in
the nineteenth century and that remain topical today. Of all
these issues, perhaps none is more central to the enterprise of
psychophysics than the matter of quantification. In a nutshell,
Murray's target article poses once again those thorny questions
that have bedeviled psychophysicists: Can sensory/perceptual
magnitudes be measured (quantified)? If so, how? And if not,
why not?
It may seem to those who are familiar with psychophysical
scaling, with its history and its ongoing controversies, that the
matter at hand constitutes another example of the proverbial
dead horse. Today, as in Fechner's time more than a century
ago, some in the scientific tribe eschew the measurement or
deny the measurability of perceptual magnitudes. And this
stance may seem justified, for, unfortunately, psychophysics has
failed so far to come up with any "measurements" with sufficient
theoretical underpinning to allay skepticism and engage the
scientific community at large into accepting them. Even so, the
matter of quantification does not go away.
Nor should it. Quantification of the mental is significant not
just in the domain of sensory-perceptual processing but
throughout psychology: in what Kurt Lewin (1936) called "val-
ence," in what economists and decision-theorists call "utility," in
what I'll loosely refer to as "valuation." In brief, valuation is the
domain of objects of desire; of the goals underlying decision-
making, including moral acts and judgments; and, ultimately, of
happiness and unhappiness - all of which have quantitative
aspects.
The treatment of mental qualities as quantities - sometimes
explicitly, often implicitly - did not appear de novo in the
nineteenth century, but has much older and deeper roots.
Plato, for example, calculated a number to represent the ratio
between the happiness of an ethical life and an unethical one;
specifically, he suggested (Republic, IX, 587) that a just ruler
(king) is 729 times as happy as a tyrant (the number three
squared, or nine, raised to the third power): "And if a per-
son . . . tells the measure of the interval which separates the
king from the tyrant in truth of pleasure, he will find him, when

the multiplication is completed, living 729 times more pleas-
antly, and the tyrant more painfully by the same interval" (Plato
1899b, p. 419).
We should not construe this numerical exercise as mere
Pythagorean - or Peloponnesian - perversion, but rather an
instance (admittedly, an extreme one) of Plato's attempt to come
to grips with such questions as how a just person should act, by
what means a person should weigh the implications of choosing
one path of action over another. To Plato, there is no disputing
calculation, and calculation demands quantification. Without a
common currency, a yardstick to hold against different options,
it would not be possible to make decisions. To choose among
alternatives is first to find what is commensurable in them, and
then to select according to which yields the greatest good.
Rarely does any objective (physical) system of measurement
give numerical representation to the various alternatives - by
and large, the numbers need to reflect valuations, which by
their very nature are psychological.
To Plato, the common currency is closely related to pleasure,
or what economists might call utility; contemporary psycho-
physicists call it hedonic value. In the Protagoras (sect. 355 ff.),
Plato (1899a) clearly metricizes good and bad in terms of plea-
sure and pain (though, I should point out, the arguments are
typically subtle, for Plato's goals were ethical, and surely he was
no proponent of a crude hedonism).
Where Plato tried his hand at theoretical psychological scal-
ing, the seventeenth-century philosopher Blaise Pascal
(1670/1958), in proposing his famous "wager," implied a formal,
normative model for decision making - quantitatively evaluat-
ing a set of expected utilities. The "wager" is to believe in (bet
on) the existence of God. To summarize the argument, Pascal
noted that all four possible outcomes (of the implicit pay-off
matrix) are entailed by the conjunction of two alternative beliefs
- belief that God does exist (B) or belief that God does not exist
(NB) - and two a priori probabilities - the probability that God
either exists (p) or does not (1-p).
By evaluating these four outcomes, Pascal sought to demon-
strate that it is "rational" to believe in God's existence. Given
any value of p > 0 (Pascal took 0.5 as indifference, but for the
sake of argument was willing to accept a very small value), if God
does not exist, it matters little whether one chooses B or NB; but
if God does exist, there is an enormous gain associated with B,
and a significant cost associated with NB. In Pascal's words, "If
you gain, you gain all; if you lose, you lose nothing. Wager, then,
without hesitation that He is" (Pense'e No. 233). Pascal speaks to
a reader who is both unprejudiced and rational, an "ideal
believer." (For a related argument, based in pragmatism, see
James 1911.) In fact, Pascal extends the argument to consider
multiple "wagers," thereby casting it in terms virtually identical
to those of modern statistical decision theory.
The quantification of qualitatively differing events is a tricky
business - a business with which physical science need have no
truck, for physical science avoids dealing with qualities. Yet
mental life is filled with phenomena that can only be charac-
terized as qualitative - and these qualitative differences are
circumvented, indeed must be circumvented, when we seek to
quantify them; for as Plato recognized, quantifying means find-
ing what is commensurable despite variation in qualitative
make-up. Quantify we must, measure we must, so as to compare
and decide. When we choose between apples and oranges, no
matter which fruit we prefer, we know that their flavors differ.
How much more difficult is the decision, how much greater the
difference in "quality," when, say, a woman must decide
whether to abort a fetus, or when a body politic or its leader
must decide whether to wage war.
Commentary/Murray. History of psychophysics
What Ekman really said
Mats Olsson,8 Kathleen Harder8 and John C. Baird"
"Department of Psychology, Stockholm University, S-106 91 Stockholm,
Sweden and "Department of Psychology, Dartmouth College, Hanover, NH
03755
Electronic mail: bjohn.c.baird@dartmouth.edu
In deriving his famous logarithmic law Fechner made two key
assumptions (Baird & Noma 1978). First, Weber's Law is
correct: The stimulus just-noticeable-difference (jnd) is a linear
function of the intensity at which it is measured. Second, the
size of the subjective (perceptual) jnd is constant for all percep-
tual magnitudes. The logarithmic "law" is theoretical, not em-
pirical, inasmuch as it is based on the untested, and in Fechner's
mind untestable, assumption that subjective jnds are constant
across the intensity scale.
As Murray relates, it has been realized for some time that the
power law would have emerged had Fechner assumed that the
subjective jnds with respect to perceptual magnitude grew
according to Weber's Law, that is, if the jnd were an increasing
linear function of subjective magnitude (Baird & Noma 1978;
Stevens 1975). This is shown in Equation 1.
A = W (1)
S. S. Stevens believed that evidence supporting this notion
had been furnished by Ekman and his colleagues at Stockholm
University, who in the late 1950s were developing analytic and
graphical methods for determining the functional relationship
between subjective variability and perceptual magnitude
(Bjdrkman 1958; 1960; Ekman 1956; 1959; Ekman & Kiinnapas
1957). In the initial work, the approach involved the use of two
psychophysical functions: The relationship between mean mag-
nitude estimates and physical intensity (Stevens functions) and
the relationship between stimulus jnds and physical intensity
(Weber functions); parameters from the two functions were
paired by stimulus attribute, though each was based on data
reported in different studies by different investigators. The
procedure consisted of marking off equal stimulus jnds on the
x-axis and graphically determining the corresponding response
jnds on the y-axis. It was then possible to plot the response jnds
against the mean magnitude estimate in order to determine a
Weber function for the response (sensation) side of the Stevens
function. A similar approach was taken earlier by Harper and
Stevens (1948), critical details of which are fully spelled out by
Gescheider (1985).
The question of interest to Ekman was whether or not this
procedure would result in a systematic, consistent relationship
between subjective jnds and perceived magnitude. The actual
results turned out to depend on the attribute (Ekman 1956). A
xmstant subjective jnd held for brightness and pitch, a linear
elationship (with an additive constant) held for lifted weights,
that is, a generalization of Weber's Law, whereas a nonlinear
relationship held for the taste of salt. Ekman seemed well aware
of the implications of these findings. He states clearly that a
generalized form of Weber's Law may be applicable to some
attributes but not others (Ekman 1956, p. 243), thus casting a
pretty long shadow over the practice of using such data in
deriving the power law. Apart from the above example of lifted
weights, we are not aware of any other studies that definitively
support Ekman's Law.
In a later paper (Ekman & Kiinnapas 1957) the functional
relationship between the subjective jnd (defined as a standard
deviation) and the mean response (fl) was formulated as a
mathematical assumption, namely,
crr = cRm + b (2)
This standard deviation, then, is an alternative definition of the
subjective jnd, different from that obtained in the earlier graphi-
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 157
Commentary/Murray: History of psychophysics
cal techniques, but still within the spirit of the Fechnerian
tradition of measuring jnds with the method of adjustment.
Equation 2 is equivalent to what Stevens subsequently called
Ekman's Law (Equation 1) only when m = 1 and b = 0. From his
reading of the Stockholm work, Stevens extracted this special
case and hence spoke of Ekman's Law because this special case,
and not the general equation actually proposed, suited his
theoretical purposes. Following Stevens's lead, the special
formulation has been referred to as Ekman's Law by researchers
ever since, a practice followed in many textbooks (e.g., Baird &
Noma 1978; Gescheider 1985), and continued by Murray in his
target article. A position more in line with the data would be to
talk of Ekman functions, meaning any functional relationship
between the standard deviation (or some other measure of
response variability) and the mean.
In summarizing his own efforts to find this relationship,
Ekman (1956) expresses doubts about the validity of his ap-
proach: "No obvious conclusions can be drawn from our survey
of experiments. We have found three different types of relations
between difference limen and position on the subjective contin-
uum . . . . It is quite clear that difference limens and scale
values ought to be determined by the same investigation" (p.
242). The same caveat applies to more recent attempts to relate
the Weber fraction to the exponent of the power function (Baird
1970; Laming 1991; Teghtsoonian 1971).
To go further, we question the ability to parcel out the
variability due to "Ekman's Law" in an experimental situation in
which subjects must match perceptions of the stimulus contin-
uum to numbers along a subjective continuum, such as in direct
scaling methods. In addition to the variability associated with
the stimulus continuum, the psychological processes underly-
ing these methods introduce substantial noise into the re-
sponses, creating response variability that swamps any variance
due to the presumed size of the jnds along the response
continuum, an outcome that Ekman and others since have
consistently shown and commented upon (e.g., Ekman & Kun-
napas 1957; Eisler & Montgomery 1974; Laming 1991; Mont-
gomery 1975). This may be why Gosta Ekman once told JCB
that while he was flattered to have a law named after him, he
would have been happier if the law were correct.
Response time based psychophysics:
An added perspective
William M. Petrusic
Department of Psychology, Carleton University, Ottawa, Ontario, Canada
K1S SB6
Electronic mall: bill-petrusic@carieton.ca
Murray's target article provides a useful and interesting exam-
ination of the development of ideas in psychophysics. To com-
plement and extend Murray's review, this commentary will try
to show that an important question in psychophysical research at
the turn of the nineteenth century concerned the proper ap-
proach to psychophysical studies. Both Cattell (1902) and
Munsterberg (1894) advocated the exclusive use of response
times with perfectly discriminated magnitudes as the proper
tools in psychophysics; they recommended forswearing confu-
sion indices for the measurement of sensation.
Cattell (1902) was clear in rejecting the jnd as a unit of
sensation: "I have argued that in most of our experiments we
have measured the error of observation, which determines the
accuracy of perception, but does not measure the intensity of
sensation" (p. 355).
Developing an indirect method for obtaining a scale of equal
sensation increments based on implicit bisection, Cattell noted:
An observer cannot decide with any certainty when the difference in
one pair of sensations is equal to the difference in another pair, but
158 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
the time it takes to perceive a difference can be measured. The
smaller the difference between two sensations, the greater is the time
to perceive it. When differences require equal times for discrimina-
tion, the discrimination is equally difficult, and the differences are
equal for consciousness . . . . We can in this way find the gray
equally difficult to discriminate from white and black, and this is the
gray which for consciousness is midway between white and black.
(p. 356)
In an experiment exploring this new bisection method, on
each trial he required each of his two observers, C (Cattell) and
X, to indicate whether the brighter of two cards (varying in
percentage of light reflected) was on the left or the right. Table 1
below provides a summary of the data reported by Cattell
(1902). Each observer provided 100 observations with each pair.
I have averaged response times over the two observers. It is
clear that these response times decrease monotonically in each
row and column, reflecting the clear dependence of response
times on the difference and ratio in physical intensity. In
addition, reflecting the compressive nature of the psychophysi-
cal function for brightness, as surmised by Cattell, the time (243
msec) to discriminate the 50% gray from black is considerably
faster than the time to distinguish this gray from white (282
msec).
Although Cattell provided a summary of his observations with
the view that response time data could be used, in principle, to
obtain a sensation scale based on indirect bisection, contempor-
ary developments in measurement and scaling (e.g., Coombs
1964; Krantz et al. 1971; Kruskal 1964; Shepard 1962) allow us to
obtain subjective scale values with the observations provided in
Table 1. Figure 1 provides a sensation scale for brightness based
on the output from the ALSCAL multidimensional scaling
algorithm (Takane et al. 1977) on the assumption that the
patterns of dependence of response time on the physical proper-
ties of the pairs to be discriminated would be evident in the pairs
not required for comparison by Cattell.
Munsterberg (1894) also asserted that response times and not
response confusions were the proper tool for the examination of
the psychophysical question, in particular, whether Weber's
Law held for visual extent and, consequently, whether Fech-
ner's enterprise was feasible:
If, then, we possessed a measure for the ease with which we can
discriminate between two stimuli, we should have a measure of their
subjective difference. In time measurements we have exactly this.
We ought to designate as equal the differences between two pairs
of intensities of stimuli, if equal times are necessary to distinguish
them. (p. 45)
In an elaborate experiment using his famous (or infamous)
chain reaction method, Munsterberg factorially varied the ratios
and the differences of the extents of pairs of lines to be discrimi-
nated. On any trial an observer viewed a pair of lines and
indicated whether the comparison line was "smaller," "equal
to," or "longer" than the standard line. In a particular series of
Table 1 (Petrusic). Cattell (1902) response time data on
brightness discrimination
75 50 25 0
100 314 282 237
75 284
50 276 243
25 251
Note: Cell entry is the mean time in msec to determine that the
row stimulus is brighter than the column stimulus. Each ob-
server provided 100 observations and cell entries were ob-
tained after averaging mean response times with the two
observers.

-2.00
0 25 50 75 100
PERCENT REFLECTANCE
Figure 1 (Petrusic). Scale values as a function of percent
reflectance obtained from a multidimensional scaling (ALSCAL)
of response times provided by Cattell (1902).
trials, for example, the standard was 5 mm and the comparison
stimuli were either 2.5, 5, or 7.5 mm, and in other series the
standard and variable stimuli arose by multiplying each of the
respective stimuli by a constant; for example, 5-10-15 mm;
7.5-15-22.5 mm; 10-20-30 mm. Overall, 36 different compari-
sons were required from each of a series of five observers and
each of these 36 comparisons was replicated 100 times with each
of the five observers. Figure 2, which provides a view of
Munsterberg's response time data, clearly shows the joint
dependence of response time on both the ratio and the differ-
ence of the physical extents. In addition to noting that the
validity of the "psycho-physic law for optic distances" was only
approximate (i.e., Weber's Law is violated because of the
dependence of response time on differences with ratios held
constant) Munsterberg concludes:
The numbers show clearly how superior the psychometric method is
to the other psycho-physic methods, for the finer analysis of the
processes in an act of discrimination [italics added for emphasis]; and
that too without regard to the fact that we are freed from the necessity
of using the just perceptible differences in away that is theoretically
questionable, (p. 51)
With the ratio of two line lengths fixed at 1.20, Henmon
(1906) and another observer each provided 320 comparisons of
line lengths differing by either 4, 3, 2, or 1 mm. For both
observers, response times clearly increased monotonically as
the differences decreased, precisely as found earlier by
Munsterberg. Curiously, Henmon remained reluctant to accept
the outcome of his own experiment. "The results of the experi-
ments on equal relative differences, as far as they go, show the
approximate validity of Weber's law. A continuation of the
experiments with a wider range of distances would be necessary
in order to obtain conclusive data" (pp. 53-54).
More recently, Petrusic and Jamieson (1979) have replicated
800
.10 4 +^
700
^Or-1.11
r-1.20v
LJ 600 -"SAME"
-"GREATER" ^
OO-"SMALLER" r-1^T>
8^*^-tr-1.5.
~-Or=2.0
400
0.0 1.0 2.0 3.0 0.0 1.0 2.0 3.0 4.0 2.S 5.0 7.S 10.012.5
DIFFERENCE IN MM
Figure 2 (Petrusic). Mean response time as a function of
physical difference in extents (mm), plotted separately for each
ratio. Data from Munsterberg (1894).
Commentary/Murray: History of psychophysics
and extended the experiment by Munsterberg (1894) by system-
atically varying both the ratios and the differences of visual
extents to be discriminated. They showed that indices based on
response times satisfied the requirements for a "positive differ-
ence structure" (Krantz et al. 1971) and that response times
were described by the following equation: RT(x,y) = F (s(x)
s(yj) + R, where RT (x,y) is the time to compare stimulus
magnitudes x and y, R is a positive constant, and s(x) denotes the
sensation activated by stimulus with magnitude x. Most signifi-
cantly, they showed that no other decision rule (e.g., a ratio of
sensations with a simple power law or a difference of sensations
with a Fechnerian psychophysical function) was acceptable, and
that in this context the psychophysical function must be a power
law with an additive constant and an exponent of less than one.
Finally, replicating and extending Cattell (1902), they used
their response times as indices of proximity and uncovered a
one-dimensional representation of visual extent, with the expo-
nent for visual extent indeed less than 1 (either .87 or .78
depending on the precise index of response time based prox-
imity used).
ACKNOWLEDGMENT
This work was supported by a grant from the Natural Sciences and
Engineering Research Council of Canada.
Unwarranted popularity of a power function
for heaviness estimates
Helen E. Ross
Department of Psychology, University of Stirling, Stirling FK9 4LA, Scotland
Electronic mall: her1@forth.stir.ac.uk
In section 1.3.4 Murray notes the enormous popularity of the
power law following its espousal by S. S. Stevens. This may well
be a case where the influence of one individual has carried
undue weight. Many contemporary textbooks stated unques-
tioningly that the psychophysical function was a power function
and failed to mention the context effects, response biasses, and
other factors that affect the form of the function (Poulton 1968;
1989). For example, Murray cites Table 5.1 in Gescheider
(1976, p. 129), which lists the exponents of the power functions
for various stimuli. However, those with exponents near 1.0 or
less than 1.0 are hard to distinguish from linear or logarithmic
functions, respectively. Exponents clearly greater than 1.0 give
more support to the power law. Most of these exponents can be
questioned, as Murray mentions for electric shock.
Another questionable exponent is that for rated heaviness,
given as 1.45. This widely quoted value is derived from Stevens
and Galanter (1957), who took the average of eight experiments
with exponents ranging from 1.13 to 2.07. The experiments
were by various authors and used various methods of ratio or
magnitude estimation. Some of these authors, particularly War-
ren and Warren (1956), pointed out that such studies may be
influenced by the size-weight illusion (SWI). The apparent
heaviness of a visible object increases with object density:
Therefore, for visible objects of constant volume but varying
mass, heaviness estimates tend to increase faster than mass.
Geldard (1972, pp. 386-87) also quoted the exponent of 1.45,
but went on to mention the bias caused by the SWI and added
"It is therefore common practice to blindfold the observer." Not
so. All of the experiments listed by Stevens and Galanter appear
to have been conducted without blindfolds and with constant
volume stimuli (though in one unpublished study by Tabak it is
reported that the variable weight was hidden from sight).
Stevens and Galanter state that they excluded the data of
Warren and Warren in which the volume of the visible objects
increased with the mass and the data of Joy (cited by Baker &
Dudek 1955) in which the objects were hidden from sight.
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 159
Commentary /Murray: History of psychophysics
These two experiments gave a near linear relation between
physical and judged weight. Poulton (1989) states in his Table
10.4 that the value of 1.45 is for seen weights; for unseen weights
(Table 4.1 and pp. 73-74) he reviews the more recent literature
and concludes that the relation is linear or logarithmic and varies
with the range and spacing of the stimuli and the type of rating
required.
Even for seen weights the exponent of 1.45 can be chal-
lenged. Stevens and Rubin (1970), Cross and Rotkin (1975), and
Rule and Curtis (1977) fitted a family of power functions to their
data, varying with several parameters including volume. If the
SWI is a logarithmic function of object density (Ross 1969;
Stevens & Rubin 1970), it would not produce a power function
when added to an underlying linear or logarithmic function for
unbiassed heaviness estimation. Cross and Rotkin argued that
the SWI should be fitted by a power function, and Rule and
Curtis that it should be an exponential decay function. The topic
remains complex.
Stevens believed that magnitude estimates were matched by
neurelectric data. This has certainly been questioned for those
modalities where neurelectric data are relevant (McKenna
1985). However, in the case ofjudged heaviness it is hard to see
what neurelectric measures could be taken, since heaviness
judgements depend upon proprioceptive information concern-
ing the manner of lifting and are entangled with many higher-
order processes (Jones 1986). Heaviness scales cannot mirror a
unitary sensory mechanism.
It is worth quoting the opinion of Stevens and Galanter (1957,
p. 386) on weight estimation: "The judgment of heaviness is . . .
easily influenced by . . . factors, such as . . . [the] size-weight
illusion, and so on. The teasing out of all these parameters would
be . . . interesting . . . if the scale of subjective weight were an
important substantive problem . . . . but . . . the substantive
outcome is usually of little interest." Nevertheless the data they
reported continue to be used to boost the validity of the power
function. Since weight judgements are so complex, it is doubtful
whether they should be used as evidence for the form of the
psychophysical law. The common use of weights as stimuli
probably stems from their convenience and historical status.
The statement that they support a power function and show an
exponent markedly greater than 1.0 probably applies only to a
subset of stimuli affected by the SWI, and this should be
stressed more than it is currently in the textbooks.
Bedrock metaphysics, fossil
fuel psychophysics
Dale A. Stout
Department of Psychology, Bishop's University, Lennoxville, Quebec,
Canada J1M 1Z7
David M urray's target article succeeds in directing our attention
to the refined arguments and research practices that Fechner
and his critics contributed to the study of psychophysics, past
and present. What Murray's "perspective" lacks, however, is a
more detailed account of the central role metaphysics played in
both Fechner's psychophysics and the ensuing philosophical
and technical debates. Indeed, Murray's selection of historical
figures appears to follow principally from their contributions to
the methodological aspects of psychophysics; those figures who
did not contribute a refinement of formulation or suggest new
methodological tactics are largely neglected in Murray's per-
spective. For example, the criticisms of Herman Lotze (1887),
James Ward (1876), and William James (1890) are not men-
tioned.
Perhaps this is justified as these authors had little impact on
psychophysics, though they all affected the way psychology was
conceived before the turn of the century. Beyond this, however,
160 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
by drawing our attention to recent accounts of similar psycho-
physical disputes, Murray's target article sets out to show us that
Fechner's speculations were sophisticated and that the ensuing
controversies were well articulated. Since this current literature
tends to be concerned less with metaphysics and more with
methodology and data-generating practices, Murray plays down
the metaphysical encrustation of psychophysics. This is unfortu-
nate because what is most admirable about nineteenth-century
psychology is that "psychologists" were "philosopher-scientists"
rather than "scientist-practitioners," as we have today. Indeed,
the nineteenth-century imagination was still set to make the
invisible visible, to craft images that "sing reality" (a phrase used
by Illich 1985, p. 11), and not to fix idols that are taken to be
reality - those metaphysically enucleated concerns of our pres-
ent psychology.
In a historical treatment I think the philosophical embedded-
ness of Fechner's ideas (and methods) and those of his critics
must be given centre stage. It was Fechner's passion to ground
his metaphysics in concrete experiment and his critics' objec-
tions to his substratum arguments that produced a fuel that
burned, a fuel fit for a veritable industry of psychological
research. Indeed, methods are used because they are under-
stood in ways that make them serve metaphysical and not just
physical ends. That is, methodologies themselves have an "in-
side " (methodology as understood by a researcher) as well as an
"outside" (methodology as understood within a tradition) aspect
(see Stout, in press).
Fechner held to a psychophysical parallelism arguing that
there was no causal relationship between the mental and physi-
cal realms though they are functionally related "as the concave
curvature of a circle viewed from its midpoint is functionally
related to the convex curvature of the same circle viewed from a
point external to it" (Fechner 1860, vol. I, p. 18). He urged in
Spinozan fashion that the law of simple proportion held within
the physical world as well as the psychical world. It was in
describing the relations between these worlds that the log-
arithmic dependency entered. That is, in drawing together
stimulus changes and the changes in sensation, the former
increased geometrically as the latter increased arithmetically.
Thus Fechner placed his law between neural processes and
mental phenomena. What he created with his psychophysical
law was a musical score for this parallelist reality.
Any perspective on psychophysics must keep this point in
sharp focus; it was the heart of Fechner's enterprise (Marshall
1990; Scheerer 1989). To seriously challenge Fechner was to
challenge his metaphysics; to challenge his metaphysics was to
effectively relocate the logarithmic dependency. To relocate
this logarithmic dependency was to question methodological
understandings, which in turn altered how psychophysics was
understood. It was clear to any nineteenth-century psychologist
that positing a relocation of the logarithmic relation was to
confront Fechner's monist metaphysics with methodologies
understood to be servants of alternative philosophical per-
spectives.
What Murray's target article lacks is any thorough discussion
of these issues, which are the very ones that have always been
the sources for psychophysical debates. Under the guise of a
metaphysical agnosticism, modern psychophysicists drift to-
ward two options: Either they adopt a cognitive analysis that
tends toward solipsism, or they opt for a physiological analysis
that still has to find a bridge back to experience. Ward
(1904/1927, p. 200) regarded the first as a "psychology of
dualism," the second as a "physics of dualism." I think his
comments on these options still hold as a guide for us: "We find
how far two halves are from making a whole; we see that the
psychology of dualism ends in shutting itself in, and the physics
of dualism in shutting itself out, by an imaginary sundering of
the one world of experience into two worlds that are alike devoid
of reality." As we can see from Murray's conclusion, he is aware
of the importance of the metaphysical/philosophical questions

in psychophysics. Yet he accepts this as representing the "slip-
pery ground" that is avoided through the study of "empirical
questions," as these represent a "terra firma." This is a false
geology.
Sensation strength: Another point of view
Robert Teghtsoonian
Department of Psychology, Smith College, Northampton, MA 01063
Electronic mall: rtex@smith.bitnet
Murray deserves much praise for his wide-ranging analysis of
the nineteenth-century antecedents of contemporary issues in
psychophysics; we owe him a debt of gratitude for helping us
better understand the important work done by the pioneers in
our field. Of the many interesting topics he has raised I have
selected just two for comment - Delboeuf's classic study of
judged greyness, and the status of the problem of measuring
sensation strength.
Murray notes that Delboeuf compared the results of his
version of the psychophysical law with the obtained data and
took comfort from the good agreement. In an extended footnote
Murray shows that Delboeuf had little grounds for pleasure with
these data since it can readily be shown that they are compatible
both with Fechner's version of a logarithmic law and with a
power law (for which Plateau was the contemporary advocate).
The point I wish to make about Murray's analysis is that it is not
really necessary to undertake the calculations made by Del-
boeuf or by Murray to show that the data are consistent with
either a logarithmic or a power law, and that this dual con-
sistency can be understood in terms of the uncertainty about
what it is that the subject is actually doing.
The important feature of all 13 sets of values for/,, J2, and/ 3 is
that they define geometric progressions. Were it not for the
fact that Delboeuf was concerned to distinguish his own model
from that of Fechner, he could have tested the adequacy of a
simple logarithmic law by using the ratio / 2 //i as a multiplier for
72 to obtain a predicted value for 73. (In fact, such "predictions"
are in excellent agreement with Delboeuf's data, perhaps better
than the fit to his own model.) It only remains to argue that the
three corresponding sensations define equal intervals on a scale
of sensation strength to allow the conclusion that there is a
logarithmic relation between sensation and degree of greyness.
The critical question, of course, is whether one can indeed
assume equal intervals on a sensation scale under the instruction
to make 72 appear "intermediate" between 7, and 73. As Murray
notes, a number of writers have drawn attention to the fact that
there need be no simple or universal relation between what the
subject is told to do and what he in fact does. Assuming that the
subject in Table 1 of the target article adjusts 73 so that S3IS2 =
S2/Slt that is, creates a geometric progression of sensations, his
results can be said to conform with a power law. (It should be
noted that this line of argument is rather different from that
proposed by Murray in Note 3 where he shows that a power law
fits Delboeuf's data even if it is assumed that the sensations in
question define equal intervals; that is the familiar demonstra-
tion that, with a low exponent, a power relation is difficult to
distinguish from one that is logarithmic.)
The problem is: How are we to know the relations among S,,
S2, and S 3 when these are not directly observable? This brings
me to the second issue on which I wish to comment, and that is
the difficulty of working with the theoretical construct "sensa-
tion strength" when there is no agreement about the defining
operations. Although everyone accepts that stimulus intensity
constitutes a major defining antecedent, there has been contin-
uing dispute about an appropriate outcome measure and
whether it is possible to specify the relation of any outcome
measure to sensation strength. Although Stevens and his col-
Commentary /Murray: History of psychophysics
leagues seemed to imply a simple linearity between magnitude
estimates and sensation strength, it has been convincingly
argued by many commentators that this is simply an assump-
tion, with little or no supporting evidence. Critics of magnitude
estimation have rightly pointed out the inconsistent and even
idiosyncratic ways in which subjects use numbers, but even an
"ideal" subject able to use numbers on a ratio scale consistently
would leave uncertainty about how those numbers relate to
sensation magnitude. But is a solution to this puzzle really
needed?
The method of cross-modal matching (CMM), mentioned
only briefly by Murray (and apparently omitted from Table 3),
seems to me to have a unique place in the context of debate
about the form of the psychophysical law. (Because matches that
involve numbers so often involve assumed relations between
those numbers and sensation strength, it will simplify the
argument to exclude such procedures from this treatment of
CMM.) Since both the target stimulus intensity (set by the
experimenter) and the matching response stimulus intensity
(selected on another continuum by the subject) are physical
values that involve no assumptions other than that of subjective
equality, the result certainly qualifies as an answer to Fechner's
question about "outer psychophysics," and for CM M the answer
has been repeatedly shown to be a power law. Indeed, it is
notable that (so far as I know) no one has ever suggested that the
power law is found for CMM only because the experimenter
induced a particular bias in his subject. Nor (so far as I know) has
it ever been reported that some subjects in CMM experiments
show power functions while others show logarithmic functions.
The results are clear, consistent, and make possible a systematic
description of a very large number of perceptual continua in
terms of the operating characteristic relating perceived magni-
tudes under the instruction of subjective matching.
Why, it may be asked, is this important discovery given such
short shrift not only by Murray but by many others concerned
with psychophysical scaling? I think the answer lies in the fact
that CMM data make no claims to measure sensation strengths.
There is nothing in such data to address the ancient questions
about ratios or differences among sensations, and it is probably
for that reason that they are of such little interest to those who
continue to see the main problem in Fechner's terms, that is,
the relation between stimulus intensity and sensation strength.
But after nearly a century and a half in which it has been
impossible to agree on defining operations for "sensation
strength" it may be time to recognize that the goal of measuring
it is a will-o'-the-wisp.
What seems to me important about CMM is that although it
makes no assumption stronger than subjective equality, it pro-
vides data that tell so much about the nature of so many
perceptual continua: The virtually exhaustive matching that
obtains for most (if not all) pairs of continua; the fact that the
matching functions generally follow a power function; the fact
that there is transitivity for the exponents of those functions; and
the likely fact that those exponents are inversely related to the
dynamic ranges for the corresponding continua (Teghtsoonian
1971). Furthermore (pace Laming 1989), if there is cross-modal
subjective equality for jnd, and the jnd is even roughly approxi-
mated by Weber's Law, then power law exponents must be
systematically related to measures of resolving power (Teght-
soonian 1974). This is a great deal to know, or even to speculate
about, in a completely testable context, without ever asking
about the relation among sensation strengths.
But if this last concept has become so compelling that few are
willing to give it up, it remains simply a matter for the Bureau of
Standards to convene a committee whose task would be to
announce a standard reference continuum, be it sound pressure
level, luminance, or our old friend the number scale, and that
would provide the operational definition of sensation strength.
So far there seems to be no experimental or theoretical solution
that seems preferable to such a resolution by fiat. Although it is
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 161
Commentary/Murray: History of psychophysics
obviously useful in many experimental contexts to have a metric
for what is called sensation strength, my point is that the choice
of a reference continuum is entirely arbitrary, and that as a
consequence the relations that obtain among "sensation
strengths" will depend on the choice that is made. Murray
characterizes such a point of view as behaviorist and unsatisfying
for most psychophysicists. Perhaps so, but my guess is that in
another 150 years the measurement of sensation strength will be
no better understood than it is now.
ACKNOWLEDGMENT
I would like to acknowledge the support of the Laboratoire de M6can-
ique et d Acoustique, CNRS, Marseille, and the director of its Psycho-
acoustics Group, Georges CaneVet; I was a visitor there while this
commentary was written. I'd also like to acknowledge the contribution
of Martha Teghtsoonian in valuable discussions of the ideas contained in who support them to clarify and defend them.
this commentary.
Psychophysics and the mind-brain problem
Michel Treisman
Department of Experimental Psychology, University of Oxford, Oxford 0X1
3UD, England
Electronic mall: treisman@vax.oxford.ac.uk
Murray's target article deals with various strands in the develop-
ment of psychophysics. The present comments are restricted to
one set of issues.
Murray's point of view largely reflects that of Stevens (1957;
1961). His presentation has the virtue of bringing to the fore
assumptions that have been implicit in the work of many
psychophysicists, including Stevens, but have not usually been
acknowledged. These assumptions constitute a philosophically
questionable interactionist basis for psychophysics. They are
represented in Murray's Figure 1 as the sequence / E > S * infinite number of triples (/'" (/"(/'))) representing the trans-
fl, where / represents the physical intensity of the stimulus,
an effect at the neural level, S sensation, and fl the overt
response. The author derives this representation from Fechner
and uses it as the basis of a classification of psychophysical
developments.
What is "sensation"? In Murray's account of Fechner's views
(sect. 1.2) "physical events in the brain were translated into
mental events of sensation'; the latter is a "mental experience'
or "conscious experience." Murray adopts and supports this
interpretation of Fechner's approach on the basis that "most
psychophysicists do want to talk about psychological experi-
ence" (sect. 2.1, para. 4).
They may well so wish, but the question is, What can usefully
be said about this topic? The unargued and uncritical adoption of
an interactionist stance fails to confront the dilemmas that that
position entails. I shall consider them briefly.
The sequence 1*E* S R represents a causal sequence. mined a scale of sensation S =f{t) having such properties that if
In the physical world, causation, the action of one event such as
the impinging of physical energy on a sense organ, to alter
another such as neural activity at an internal locus, is subject to
physical laws. Thus, when a racket strikes a ball, energy is
transferred and entropy increases but the total energy of the
system is conserved. Such laws apply to the neural stage, or
series of neural stages, that lie between I and E. What about the
sequences S and S fl?If a given value of causes a given S'3/S'4, that is, equal contrast would appear to be determined
strength of S to be produced, and this value of S initiates a
corresponding response fl, and these causal events are gov-
erned by the laws of physics, such as the law of conservation of
energy, no problem arises: except that if S is an event governed
by the laws of physics, then it is a physical event, one that
happens to occur in the brain. Inasmuch as the physical compo-
nents of the brain are neurons it is a neural event, and thus S is a
stage of the same sort as ; and so in explaining how / produces
162 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
fl, both E and S refer to intervening neural stages and one or
other of the symbols is redundant.
We have arrived on one horn of a dilemma, in which the
special status of S disappears, and to avoid the discomfort this
causes we must switch to the other horn. That is, to preserve the
special status of S in this context we must claim that S is not a
physical event governed by physical laws but a mental event not
governed by such laws. What can this mean? That when E acts to
determine S, the law of conservation of energy does not apply:
Energy disappears from the physical world? Or that the neural
acts that determine S do not dissipate energy? That when S acts
to produce fl it transfers energy from a mental into the physical
world? Or that S produces physical effects, changes in the states
and actions of neurons, without transferring energy to them?
These are very difficult positions, and I shall leave it to those
It is clear that we cannot simply insert "S" into a causal
sequence without acknowledging the problems and answering
the questions that the assumption of an intervening interactive
mental stage raises. But even if we were to assume for the sake of
argument that these assumptions were not problematic, that we
can postulate the sequence / * E * S *flwithout explanations
being required, there would still be serious difficulties in any
attempt to determine "inner psychophysical" functional rela-
tions that are not acknowledged by the article.
I have argued previously that Stevens's (1957; 1961) attempt
to interpret direct scaling results as evidence for a "law" relating
stimulus magnitudes to sensation intensities is flawed (Treisman
1961; 1962; 1963a; 1963b; 1964a; 1964c; 1985; 1989); that if we
assume a sequence such as / * E * S * fl, and have
experimental observations on the relation between / and R,
these data cannot uniquely specify each stage in this sequence.
If we determine the relation fl =j{I) experimentally, we cannot
arbitrarily assign this relation to one or a subset of the intermedi-
ary stages above, such as S, or / * S, however much it
might please us to do so. One can replace the relation/with an
formations from / to , from to S, and from S to fl (i.e., = / '
(7), S=f (E), fl =f"{S)), where the functions are so chosen that
any one of these triples would be equivalent to fl = f{l). Thus,
from the observed relationship between R and / we cannot
derive a unique relation between S and /, or between S and E.
This argument is ignored, not answered, by Murray's treat-
ment. Let me illustrate the application of the argument to his
discussion of lightness contrast. Consider a range of luminances
in which we perceive the same contrast between two surfaces
when the same objective ratio holds between the light reflected
by each surface, whatever the absolute luminance level. The
author of the target article is concerned whether this is deter-
mined by our perceiving a constant degree of contrast for
constant ratios between sensations or for constant differences
between sensations. But is this a substantive issue?
Let us suppose that in such a case we have somehow deter-
the contrast between stimulus intensities 7t and 72, on the one
hand, and stimulus intensities 73 and 74, on the other, are judged
to be the same, then S, - S2 is equal to S3 - S4. That is, Murray
would claim that equal contrast is determined by equal sensa-
tion differences. But it is fully open to us to substitute for S a new
scale of sensation defined as S' = es = ef(l>, and on this scale it
would be equally true that for equal contrast S'JS'z is equal to
by equal sensation ratios. Whatever experimental predictions
may be derived from the model containing S can equally well be
derived from the parallel model containing S'.
Thus there is no substantive problem requiring us to decide
whether the perception of constant contrast for the same objec-
tive ratios between luminances is determined by our perceiving
constant contrast for constant ratios between sensations or for
constant differences between sensations. The answer depends

on an arbitrarily chosen scaling convention. We can scale the
mechanisms determining our perception of constant contrast so
as to give either answer, just as the intensity of a tone can be
measured in decibels or power units. Within the scope of this
kind of experiment, the information does not exist to give one or
other scale a special status (Treisman 1961; 1962; 1963a; 1963b;
1964a; 1964c; 1985; 1989).
The analysis of sensations as the foundation
of all sciences
J. van Brakel
Department of Philosophy, University of Utrecht, 3508 TC Utrecht,
Netherlands
Electronic mail: brakel@phil.ruu.nl
In 1932 a committee of the British Association for the Advance-
ment of Science was appointed to report on "the possibility of
Quantitative Estimates of Sensory Events." Eight years later
(Ferguson 1940):
The Committee . . . found it possible . . . to state their conclusions
in terms that all of them could accept; but in doing so they . . . had to
ignore the wording of their terms of reference. . . . After examining
[the various] points of view [concerning the possibility, etc.] the
Committee [were] satisfied that no practicable amount of discussion
would enable them to express an agreed opinion. . . .
The same might be said of the discussion 50 years earlier
between Fechner and von Kries. And again the same could be
said of the commentaries on Krueger (1989a) and on Murray's
target article.
My comments on Murray's target article will be restricted to
section 1.5.3 and will deal with the views of von Kries and Mach.
I will suggest that Mach's views might provide a "novel" per-
spective on psychophysics.
Von Kries's (1882) argument is that on a practical level
Fechner's jnd scales (or any other quantitative scale for sensa-
tions) may be useful for some purpose or other, but on a
theoretical level the idea of a sensation scale in the same sense as
a scale for length, mass, and time is meaningless, because the
very idea of equality of sensation difference (at different points
along the scale) is empty: "die Gleichsetzung zweier ver-
schiedener Empfindungszuwuchse . . . hat gar keinen Sinn"
(p. 274). For any particular type of sensation, not only does a jnd
depend on "surrounding" sensations, the part of the sense organ
used, the history of exposure, and so on, but the whole idea of
looking for units of sensations is meaningless. That people are
often good at "producing' ratio or interval scales in experiments,
just means that they have been trained through survival or
otherwise to be good measurers of physical magnitudes. (I have
been told that there are people who are extremely good at
estimating the average water content of large piles of fresh
natural rubber. Do they have "average water content of large
piles of fresh natural rubber-sensations"?)
Basically, von Kries's message is: Use any scale that serves
your needs, but don't think there's a reality behind it (such as
the true metric structure of sensation space). Note that this gives
a somewhat different emphasis from Murray's account. He
seems to suggest that if von Kries is right, there's no room for
psychophysics. However, the only thing von Kries opposes is
Fechner's appeal to the correct unit of sensations. He concludes
(p. 277) "dass wir es mit einer willkiirlichen Festzetzung von
Grossenbezeichnungen zu thun haben, welche zweckmassig
oder unzweckmassig sein kann, aber nicht richtig oder falsch."
Similarly, I couldn't place Murray's suggestion, on Fechner's
behalf but against von Kries, that "astronomers can measure
star-brightnesses" and that "Delboeuf's subjects could indicate
[an intermediate] greyness." Von Kries (1882) discusses both
examples in detail, but concludes they merely show that we are
Commentary /Murray: History of psychophysics
good at using our senses to make guesses at the objective
properties of the world. For example, he argues that brightness
tends to be judged in terms of proportions (ratios), because it is
proportional to background illumination, changing atmospheric
absorption (in the case of stars), the size of the pupils, and so on.
Nevertheless, von Kries is unsympathetic to psychophysics in
the sense that sensations themselves cannot be measured in the
same way as physical length, time, and mass. However, the
significance of the difference (in particular, its negative appeal)
only makes sense if the priority of physics is taken for granted. It
is here that Maeh offers an alternative. (I am not endorsing this
alternative - just suggesting that it might throw light on what
the purpose of psychophysics could be.) His view can be
summarized as follows:
(1) No physical (or any other) measurement is possible with-
out measurement of sensations. The measurement of sensations
provides the basis for the unity of knowledge, grounded in the
manifest (daily, folk, common sense) experience of the world.
There is therefore no ground to worry about psychophysics
losing out on physics.
(2) The sense of measurement intended, however, is that of
ordinal measurement. Interval or ratio scales for sensations (or
judgments for that matter), as well as correlations that look like
psychophysical laws, are artefacts of data-gathering procedures
or the result of entrenched conventions.
How should we understand Mach's view that the most basic
form of measurement is the measurement of sensations?
(Though not a literal quote, I think it captures Mach's spirit.)
The first thing to do is to forget about the difference between the
things mentioned in the left and right column of the table below:
sense perception of time clock time
sense of space (seen length, area, volume) length, area, volume
felt force/pressure/pull/push physical force
felt warmth temperature
(musical) pitch frequency (of longitudinal waves)
experience of brightness luminosity, quantity of light
According to Mach (in the interpretation I'm offering) the true
nature of the world consists of elements that correspond to the
rows in the table above. But there's only one column. Making
the distinction between, say, felt warmth and temperature is a
later (human) addition. Having made the distinction it turns out
that the two corresponding entities in any row are in ordinal
correspondence. This fact is the basis of all measurement. If
there were no prior ordinal experience (on the level of the
elements), there would be no metrical science. This ordinal
space of (pre)sensation elements is, as it were, the metaphysics
on which both psychology and physics are built (cf. Bradley
1971, p. 68). Failure of transitivity and so on should be under-
stood in terms of more than one row being actual - an experi-
ment "about one row" always abstracts, if only from a
background.
When the physical column has been constructed, empirical
laws can be found that warrant the use of ratio scales. This is not
the case for the sensation column (here Mach agrees with von
Kries). However, it's easy to overestimate, as von Kries does,
the metaphysical significance of this (as distinct from the practi-
cal relevance). Mach's advice to psychophysics is, I suggest:
Analyse sensations - it's the most fundamental science. But
don't worry about scaling (or the "true" form of the psychophysi-
cal law if you think that is something different): Any perspicuous
and insightful presentation of your research will do.
As I do not expect everyone to be convinced by Mach's lack of
deference for physics, I therefore add two examples which may
help to realise "the danger of overestimating the instruments,
with which we are so constantly employed, or even of regarding
them as the objective point of science" (Mach 1960, p. 609).
Temperature is often used to illustrate interval scales or
difference measurement. Stevens and others use it as an exam-
ple to illustrate the "evolutionary" development from nominal to
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 163
Commentary /Murray: History of psychophysics
ordinal to interval to ratio scales (degrees Celsius and Fahren-
heit would be interval scales; the Kelvin scale, with an absolute
zero, would be a ratio scale). Temperature is of interest because
Mach explicitly draws an analogy between sensation scales and
temperature scales. After arguing that, at least for brightness, a
proportional psychophysical law makes more (biological) sense
than Fechner's Law, he writes (Mach 1919, p. 67n - quoted in
part by Murray):
Genau genommen, halte ich auch den Ausdruck Proportionalitat fur
unzutreffend, da von einer eigentlichen Messung der Empfindung
doch nicht die Rede sein kann, sondern hochstens von einer genauen
Charakterisierung und Inventarisierung durch Zahlen. Vgl. Prinzi-
pien der Warmelehre, S.56, was uber die Bezeichnung der War-
mezustande gesagt ist.
And there h e writes (Ellis 1968, p . 196):
Temperature . . . is nothing but a characterization, or identification
[Kennzeichnung], of the heat-states by a number. This temperature
number has the property merely of an index number, by which we
can recognize the same heat-state again, and, if necessary, locate and
reproduce it. At the same time, this number allows us to recognize
the order in which the signified heat-states follow upon one another,
and between which other states a given state lies.
Part of the problem with the temperature scale is the fact that
the choice of the thermometric substance is arbitrary. This is the
same kind of arbitrariness that is emphasised by von Kries.
According to a realistic (materialistic) world view the only way to
transcend this arbitrariness is to ground scales in (metric)
properties of natural kinds. This can be clearly seen in the
development of the choice of units or standards for physical
magnitudes (van Brakel 1990). In this sense, Fechner's intuition
of searching for natural units of sensation is better than that of
Stevens, who appeals to the purely pragmatic character of the
choice of a unit. But in choosing an arbitrary physical unit as
the referent of the sensation unit, Stevens in fact removes the
sensation level altogether.
So Mach's suggestion is: Any sensation scale is like a tempera-
ture scale, whereas a temperature scale is, contrary to our
"secure attunements," nothing more than an ordinal scale. I do
not have space tb go through Mach's arguments (and the useful
elaborations of Ellis 1968), but for a start consider the following:
Assume temperature is measured using T = log V, where V is
the volume of a perfect gas. Would this "scale transformation" be
wrong? No. Would it make any difference to physics? Yes, here
and there an exponent, but nothing substantial, would change.
Did the arbitrariness of the ordinary temperature scales hamper
the development of physics? Not that I know of. Nowadays,
physicists happily refer to "negative, infinite and hotter than
infinite temperatures" (Ehrlich 1982), where temperature is still
measured on a "ratio" scale with a "natural" zero.
As a second example of the fragility of our secure attune-
ments, consider the measurement of length and the following
alternative concatenation procedure (Ellis 1968; van Brakel &
van der Peut 1979): Instead of putting two rods one after the
other along a straight line (call this operation +), concatenate
them by forming a right angle and call the distance along the
hypotenuse the (physical) sum of the length of the two rods (call
this operation x). Both operations satisfy the same axioms (of
extensive measurement); both scales are additive. If the same
unit is chosen, the relation between the two scales is such that
the number assigned on the "normal" +-scale is the square root
of the number assigned on the "abnormal" x-scale. Although,
allegedly, the only arbitrary choice for a ratio scale is its unit, in
the example just given, the same unit can "support" two differ-
ent scales, which are theoretically equally adequate. It may be
more practical to use the +-scale as it leads to a (mathematically)
"simpler" sort of physics, though it wouldn't change anything to
the substance of physics if we used the alternative scale.
164 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
On the construction of
psychophysical reality
Mark Wagner
Department of Psychology, Wagner College, Staten Island, NY 10301
David Murray should be congratulated for presenting an inter-
esting view of the history of psychophysics. His target article
demonstrates that in spite of over a century of work in the field a
number of issues remain controversial. I would like to make a
few comments about three issues Murray focuses on.
First, are sensations measurable? Von Kries was not the only
early psychologist to question the measurability of sensation.
For example, William James (1890) had a low opinion of Fech-
ner's work. In his words: "The whole outcome of the discussion,
so far as Fechner's theories are concerned, is indeed nil." (p.
548). Beyond feeling that the whole enterprise was boring and
without practical importance ("The whole notion of measuring
sensations numerically, remains in short a mere mathematical
speculation about possibilities, which has never been applied to
practice" (p. 539), James had three basic objections:
(1) James cites an article by Wundt (1863) that states (in
modern terminology) that it is impossible to establish more than
an ordinal scale for sensory magnitude since we do not know
what it means to say or how objectively to show that one
stimulus is subjectively twice (or * times) bigger than another.
(2) James followed this by arguing that Fechner's assumption
that just noticeable differences (jnds) are all subjectively the
same size is "entirely arbitrary." Small jnds may be perceptually
small as well, while large ones may be perceptually large. To
James, Weber's Law was more or less a fact, whereas Fechner's
thoughts about subjective reality were mere speculation.
(3) James felt that our sensations of magnitude are "indivisible
units," not just a piling up (as with summated jnds) of smaller
sensations. Like Stevens (1957), James saw jnds as measures of
resolution, and thought it was not sensible to combine them. In
fact, intense sensations are often qualitatively different from
their less intense cousins. For example, the sensation of red is
qualitatively different from a combination of light pinks.
Titchener (1896) presented essentially the same argument
(without giving credit to James), while adding another illustra-
tion. A subject is presented with two glasses of water, one with a
little sugar in it and a second that is thick and syrupy. Titchener
notes that "there is no hint - if you think only of the tastes
themselves, and forget about the sugar - that the strong sweet
contains the weak. And it is impossible to imagine a third sweet
which, if added to the weak, shall give the strong: the adding of
sensation to sensation is meaningless, a task you cannot lay hold
of" (p. 203).
My own opinion concerning the measurability of sensation is
more charitable than James's and Titchener's. I believe that
during ordinary awareness, people do not attend to the magni-
tudes of their sensory experience or attach numbers to it. Yet,
when placed in a situation like a psychology experiment requir-
ing them to make such judgments, people are cognitively
flexible enough to construct a scale relatively consistent with
their experience and the situation's demands.
This brings us to Murray's second issue, "outer psycho-
physics," which concerns determining the psychophysical law
relating the observable physical stimulus and the psychological
response. Murray points out that this issue has remained stub-
bornly unresolved for more than a century. I believe that the
reason for this lack of progress is that the issue is actually
unresolvable.
To my mind, the issue is similar to what we find in size-
distance perception (Baird & Wagner 1991). The psychophysi-
cal functions relating perceived size to physical size as a function
of distance are utterly different when subjects are given projec-
tive instructions (which ask subjects to take on an artists-eye
view), apparent instructions (which ask subjects to say how big

the stimulus "looks"), and objective instructions (which ask
subjects to take into account distance to determine the objective
size of the stimulus as measured by a ruler). The different
psychophysical functions arise because the instructions induce
subjects to look at their subjective experience in different ways.
These different perspectives each give a slightly different mean-
ing to the word "size."
Similarly, the different psychophysical methods subtly ask
subjects to look at their subjective experience in ways that give
slightly different meanings to the word "magnitude." Some
instructions ask subjects either explicitly or implicitly to pre-
serve differences, and subjects dutifully construct a scale, a way
of looking at their experience, which more or less preserves
differences. Some instructions ask subjects to preserve ratios,
and subjects dutifully construct a scale, a way of looking at their
experience, which more or less preserves ratios. Undoubtedly,
subjects are cognitively flexible enough to produce other psy-
chophysical functions if an experimenter cleverly presents the
right instructions.
Finally, I would like to discuss briefly Murray's third issue,
the possibility of "inner psychophysics." I believe inner psycho-
physics died the death it deserved for attempting the impos-
sible: resolving the mind-body problem empirically. (It's a
stretch to say signal detection theory derived from inner psycho-
physics. Noise, in SDT, refers to the mathematical construct of
variability and is only peripherally related to neural activity.)
Although one could trace neural firing patterns as a function of
stimulus intensity upon the nervous system, how could one ever
know which collection of neurons represents the crossover point
from brain to mind? Is it the brain as a whole, some section of the
cortex, the thalamus and something else, or maybe Descartes'
pineal gland? Higher in the nervous system, neural activity will
be a mixture of thoughts and feelings, as well as sensations. At
the critical point, it may be difficult to know what neural activity
produces sensations of magnitude apart from other activity.
Finally, as in outer psychophysics, if we can't know with cer-
tainty what the "true" magnitude of a sensation is, we can't
derive the "true" relationship between neural activity and
sensation either.
Looking backward: Progress
in outer psychophysics
David J. Weiss
Department of Psychology, California State University, Los Angeles, CA
90032
Murray's presentation of Fechnerian outer psychophysics
makes it clear that Fechner did not view the form of the
psychophysical law as an empirical issue. Recent efforts by
mathematically oriented colleagues - perhaps most prominent
are Luce and his associates (e.g., Luce 1959; Luce & Narens
1987) and Norwich (1987) - follow the tradition of deriving the
law from basic principles. This perspective is based on the belief
that there must be a general function relating stimulus to
sensation.
The dominant alternative view is that each stimulus domain
(and possibly each judgmental task - see Schneider 1989) calls
for independent investigation and validation of the obtained
scale. From this perspective, whether a common law holds
across continua is a question for the data to decide. Stevens and
his descendants (I trust Krueger [1989a] will not object to my
characterization of his lineage) have presented themselves as
empiricists, and have allowed an immense body of observations
to persuade them that there is a unified law. Others argue that
although the data are reliable, they cannot be precise enough to
Commentary/Murray: History of psychophysics
distinguish among contending functions which make essentially
equivalent predictions. A different, nonempirical, argument
against a universal law focuses on the arbitrary measure of the
physical stimulus (Myers 1982; Weiss 1981; 1989).
Another nineteenth-century issue brought out by Murray is
also still with us. Fechner criticized the introspective validity of
Hering's difference sensations. Validity is the vexing concern in
empirical attempts to determine sensation values. Despite a
fairly sophisticated typology of the concept (e.g., Nunnally
1970), validity ultimately seems to be in the eye of the beholder.
Progress might be made by viewing a response as a judgment
rather than as the observer's reading a number from an internal
meter. Then, by modeling the judgment, the researcher can
attempt to validate a view of the judgmental process.
This cognitive emphasis characterizes the functional mea-
surement approach to the issues of psychophysics (Anderson
1970). Sensation scale values fall out as a byproduct of successful
modeling. A characteristic of functional measurement analysis is
that it calls for complex judgments, in which two or more stimuli
are integrated by the observer, presumably in accord with the
hypothesized model. A major advantage of this approach is that
context effects are a positive force. They enrich the model,
providing additional leverage for analysis (Anderson 1975; Birn-
baum 1974) rather than obscuring pure estimates of sensation as
the earlier tradition held.
The nineteenth-century technique of equisection offers a
beautiful illustration of an integrative judgmental task. Func-
tional measurement affords a validation scheme for equisection
and allows the derivation of psychophysical scales (Weiss 1975)
in a way I like to think would have pleased Fechner. By allowing
for unequal weighting of the integrated components, the anal-
ysis avoids the assumption (also made by Garner 1954) that the
subject divides the interval in half. Finally, Fechner would
surely have been pleased at the way methodological and analytic
procedures developed using psychophysical tasks that have
been found applicable in other domains (Anderson 1982).
A perspective for viewing the present
of psychophysics
Paul Whittle
Department of Experimental Psychology, University of Cambridge,
Cambridge CB2 3EB, England
Electronic mail: pwi09@phoenix.cambridge.ac.uk
David Murray-has given us a salutary reminder that Fechner
and others were much smarter than our stereotypes of them. I
should like briefly to discuss two points: (1) the meaning of the
word "psychophysics"; (2) the metamorphosis of discoveries into
tools. This has happened to thresholds and their variability. It is
time for it to happen to Fechner's Integration.
Although he expands our historical horizons in reminding us
of Fechner's concern with inner psychophysics, when Murray
comes to discuss contemporary psychophysics, he uses the word
in a sense that will strike many people who think of themselves
as doing psychophysics as unduly restricted. He includes, on
the one hand, sensory scaling, and on the other, psychophysical
method. This shows piety towards Fechner, but perhaps does
him a disservice in remaining too fixated on the problems he
raised, rather than building on his work in order to go beyond it.
The word "psychophysics" is now very commonly used in a
broader sense, to mean something like "the study of perception
by means of the behavioural responses of intact organisms"
(where the responses of course include speech). This sense is
exemplified by the hundreds of studies (for the most part, using
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 165
Response/Murray: History of psychophysics
sinusoidal gratings) that are summarised by Graham (1989).
Psychophysical "laws" or the measurability of sensation (or
lightness constancy) are nowhere discussed in that book. This
implicit acceptance of von Kries's position has not, as Fechner
feared, led to the abandonment of psychophysics, indeed quite
the opposite. It has produced a most impressive "body of
measurements to contemplate."
We can see in this work transitions from discovery to method
that are not only a recurrent feature of science but a sign of its
health. Psychophysical methods are prominent, but they are
used only as methods. Questions of the reality of thresholds, or
of the constancy of variance of the signal-plus-noise distribution,
which once seemed to be substantive generalisations worth
fighting over, are now seen to be local questions to be answered
in particular situations not for their own sake, but en route to
something of more current interest. We now have a body of
psychophysical method and theory that provides the analytical
tools to do this (e.g., Macmillan & Creelman 1991).
I have recently argued (Whittle, in press) that "Fechner's
integration" - the idea that we can measure sensation strength
by summing jnds - should also be seen as the basis for a method
rather than as a substantive hypothesis. In the situation I
studied there was good agreement between jnds as measured by
a two-alternative forced-choice task and step size in an equal-
subjective-interval series. I do not think, however, that we
should take this simply as vindicating Fechner, and wait with
bated breath for someone to produce a counterexample. We
should rather take it as showing that the idea holds in one
situation (and note the methods needed to show this), although
it may well not hold in others. This would not be a stalemate but
a pointer to interesting differences between the situations. For
example, it would suggest that in the second case different cues
were used for discrimination and for scaling. We need more
techniques for studying perception of suprathreshold stimuli;
Fechner's idea points to one that has the very desirable property
of allowing us to evaluate the degree of convergence between
different methods.
The same argument can be made for another topic Murray
discusses, the constancy of lightness if luminance ratios are
constant. This holds quite precisely in some situations but not at
all in others. The difference can be seen not as the failure of a
promising generalisation but as an important pointer to differ-
ences between subjects' interpretations of the two situations in
terms of surface colour and lighting (Whittle 1991).
In these instances the old questions are not settled so much as
put into perspective. We see with hindsight that their impor-
tance was to spur us to develop methods for doing this. As
generalisations, they become of "merely academic" interest.
Author's Response
The place of psychophysics in the history
of sensory science
David J. Murray
Department of Psychology, Queen's University, Kingston, Ontario, Canada
K7L3N6
Electronic mall: murrayd@qucdn.queensu.ca
When I first read the assembled commentaries, I was
struck by how many added to and amplified the argument
in the target article, so that a reader who started with the
target article and then read the commentaries in order
might appreciate an opportunity to have the gist of the
commentaries reorganized within the "perspective" of
166 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
the target article. I shall attempt to do this here by
presenting the framework of the target article again but
this time rounded out by discussion of the contributions
in the commentaries and by some of my own addenda.
Actual criticisms of my remarks in the target article will be
dealt with in the appropriate place. The plan of this
response will first involve some integration of various
commentaries concerned with the antecedents and back-
ground of Fechner's Elements of psychophysics (1860/
1964). Then we shall cover the same ground as the target
article but accentuating certain points, particularly the
issue of the logarithmic law and that of sensation measure-
ment. We will conclude by elaborating on two matters
raised by several commentators: the need for psycho-
physicists to take context effects into account and the
suggestion that they transfer their focus from measure-
ments of sensation magnitude to measurements concern-
ing the duration of psychophysical decisions.
R1. Historical and philosophical precursors
of Fechner's psychophysics
R1.1. Fechner's predecessors. Exactly when mathema-
tics first entered psychology is a matter of debate. Marks
informs us that Plato tried to estimate how much happier
ethical people were compared with unethical people, a
question itself open to question ever since Freud taught
us about the dangers of a dominating superego. Marks
also points out that both Plato and Pascal described
situations in which the subject basically carries out "intu-
itive statistics" to make decisions as to what to do or
whether to believe in God. In the latter case, I again refer
to depth psychology, which could claim that no matter
how rational you may wish to be in deciding whether or
not to believe in God, if the idea of "God" was put into
your head as a child you will probably never really escape
"believing" in God no matter how hard you try. Despite
the fact that many of our decisions are hopelessly biassed
by our emotional needs and by our childhood learning,
intellectuals since Plato and Pascal have persisted in
trying to bring numbers into psychology. Thomasius in
1691 asked his readers to rate personality characteristics
(e.g., greed for money) on a scale of 1 to 60 (see
McReynolds & Ludwig 1977; Ramul 1963, p. 657);
Hutcheson in 1725 invented a "moral algebra" in which he
basically elaborated on Pascal's notion that moral good
could be ascribed a numerical rating (see Brooks & Aalto
1981); Wolff (1731/1968) first used the word "psycho-
metrics" as a name for a proposed branch of science in
which numbers could be ascribed to degrees of "pleasure"
and "displeasure" (see Metraux 1983; Ramul 1960, p.
256); and, as Gregson stresses, the first full-scale attempt
to bring mathematics into psychology was made by Her-
bart in individual articles (e.g., Herbart 1812/1889), in his
book Textbook in psychology (1816/1891), and in his very
large work whose title translates as Psychology as a
science newly founded on experience, metaphysics, and
mathematics (1824/1892). The historical background to
Herbart's mathematization of psychology is described in
detail by Leary (1980), who notes direct influences on
Herbart's thinking of the writings of Leibniz, Kant, and
others.
I am grateful to Gregson for finding Herbart's article of
1812, whose title translates as "Psychological investiga-

tion concerning the strength of a given presentation as a
(unction of its duration." The German word Vorstellung,
which literally means "a placing before," is translated
here as "presentation," following G. F. Stout (1888);
others have translated it as "concept" or "idea." In his
model, Herbart conceived that the various presentations
that are in consciousness at a given time all exerted
inhibition on each other, so that after an interval many of
them would be pushed below the threshold of conscious-
ness; he speculated that a differential equation described
how the amount pushed below the threshold varied as a
function of time and the amount above the threshold.
There were also other assumptions concerning how the
inhibition of a given presentation varied with the strength
of the competing presentations. In particular, it was
claimed that two strong presentations could almost totally
inhibit a third if it was less strong than the other two. At
the same time, other postulates determine how one
presentation could fuse with another or help another up
into consciousness. I cannot summarize the rest here; I
once thought of trying to summarize his Psychology as a
science, but found Herbart's arguments so condensed
that I am now convinced only a literal translation, or
extended summary, by a fluent German speaker with a
strong background in mathematics will do. (An excerpt
from the beginning of the work is translated in Shipley
1961.) However, as Gregson points out, Drobisch
(1850/1972) published a book on "mathematical psychol-
ogy" 10 years before Fechner's Elements. This book is
largely an attempt to make Herbart's mathematics more
palatable to the general reader but still includes pages
filled with equations; and Gregson now gives us his
opinion that Herbart's formulation is "coherent and ties
together what later fell apart."
One of the reasons Herbart's work was later ignored
was the strong influence of Wundt who, in his epochal
preface to his Beitrage zur Theorie der Sinneswahrneh-
mung [Contributions to the theory of sense-perception]
(1862/1961), wrote a manifesto that laid the foundation for
an institutionalized science of psychology. Wundt ad-
mired Herbart for his ambition in trying to found a
scientific psychology, but he thought of his mathematics
as fruitless and as resting on errors. Wundt spelled some
of these out on pp. 796-99 of the first edition of his
Grundzuge der physiologischen Psychologie [Principles
of physiological psychology] (1874), repeating his com-
ments unchanged in later editions. For Wundt, one of
Herbart's assumptions concerning the way the inhibition
arising from two presentations could almost overwhelm a
third presentation seemed unreasonable; moreover, in
Wundt's system spontaneous processes in the course of
"apperception" could determine the clarity of presenta-
tions in the early phases of perception. Wundt claimed
that Herbart's system did not make enough of these
spontaneous processes - an account of consciousness in
terms of competing presentations seemed sufficient in
Herbart's system. On the other hand, Ebbinghaus
(1885/1964) did an experiment aimed directly at testing
one of the consequences of Herbart's theory, which was
that if a few items enter consciousness in order A, B, C,
D . . . there will be a strong association between A and
B, a less strong association between A and C, and a weak
association between A and D.
Gregson wishes to revive Herbart's theory, partic-
Responsel'Murray: History of psychophysics
ularly in view of the fact that numerical methods carried
out by computer might answer some questions raised by
Herbart that do not have known analytic solutions. I can
only applaud Gregson's optimism, but we await a good
English translation or summary of Herbart's mathemati-
cal theory. There is a little in the translation of his popular
Textbook in psychology (1816/1891), but: it gives only a
taste of the immensity of Herbart's model; the summary
of Herbart's ideas by G. F. Stout (1888) is nonmathemati-
cal, but otherwise the best account available; he stresses
how Herbart had his own theory of apperception and was
a pioneer of the notion that the ego, or self, acts as a
unifying power in cognition.
Fechner's work was also preceded by a tradition of
experimentation on the special senses. Boring (1942) and
Ramul (I960; 1963) have shown how various researchers
in the eighteenth century explored visual acuity and the
duration of visual aftereffects; J. G. Kriiger in 1746 postu-
lated that the intensity of sensation increases in propor-
tion to the increase of a corresponding stimulus (see
Ramul 1960, pp. 261-62); I. M. Chaldenius in 1750
calculated how long it took for a sensory perception,
memory image, or fantasy image to become distinct in
consciousness (see Ramul 1963, pp. 655-66); Bouguer in
1760 anticipated Weber's Law in describing the results of
an experiment on just noticeable differences in brightness
(see Boring 1942, pp. 136-37); and now we have new
information, here summarized by Griisser, about the
pioneering eighteenth-century work of Tobias Mayer
(1754b; 1755; his work on acuity is also briefly mentioned
by Ramul 1963, p. 653). We are all interested to learn that
Mayer found visual acuity could be related to the overall
illumination of the target material by a power function,
although I am not sure that this directly anticipated
Plateau or Stevens because Mayer's dependent variable
was visual acuity and Plateau and Stevens's dependent
variable was sensation strength. In carrying out experi-
ments on sensation, early nineteenth-century re-
searchers invented psychophysical methods before Fech-
ner (e.g., Delezenne's work in 1827, as Gregson points
out); and Weber's research on differential thresholds with
weights placed in the hands or lifted, as well as his
research on the two-point threshold, clearly involved
forerunners of Fechner's methods; both Borg and Greg-
son believe that Weber's contributions have been
underestimated.
From a more general point of view, it can be claimed, as
Turner (1982) has, that psychology got its foot into the
door to respectability by concentrating, in its early
phases, on sensory science, which, in the hands of Helm-
holtz in particular, had high status in Germany. In fact, as
Heidelberger (1993a) stresses, psychologists may have
been persuaded to abandon the philosophy departments
because of psychologists' interest in the measurement of
sensations; and the laboratories of Wundt at Leipzig,
Ebbinghaus at Berlin, and G. E. Miiller at Gottingen
provided refuges for those who wanted psychology to
abandon its armchair tradition.
R.1.2. Fechner's own achievement. It is pointed out by
both Gundlach and D. Stout that Fechner's view of the
physical and the mental as representing two aspects of the
same substance may be termed Spinozistic; in Spinoza's
philosophy what had seemed a definite dichotomy in
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 167
Response/Murray: History of psychophysics
Descartes' philosophy - the division between the spa-
tially extended (matter) and the nonextended (mind) -
was argued to be faulty insofar as God (normally thought
of as nonextended) could also be extended. Fechner's
system rarely mentions God in the context of his psycho-
physics, though God is discussed elsewhere in his writ-
ings (see Lowrie, 1946, for translations of Fechner on
religion); he thought of religion as the highest of human
achievements, higher than art (Fechner 1876, vol. 2, p.
160); and once Spinoza had questioned the need for a
dichotomy between the extended and the nonextended,
the way was paved for Fechner's replacement of Carte-
sian dualism by his own form of monism. Nevertheless,
there are different classes of monism; the view that some
kinds of matter (notably brain tissue) are associated with
the subjective experience of "mind" because "mind" is an
epiphenomenon of brain activity is a modern kind of
monism; Fechner's monism arose from a tradition of
philosophy, going back to some extent to Spinoza, but to a
greater extent to Leibniz, according to which all matter
has a mind-or-consciousness-aspect to it, with some kinds
of matter (such as human brain tissue) being associated
with the kind of consciousness we are familiar with. As
Gigerenzer indicates, Fechner (1848) argued strongly
that plants have a kind of consciousness too, but, because
they do not have a nervous system (brain tissue), the kind
of consciousness they have might never be experienced or
understood by humans. When Leibniz put forward the
idea that all matter is associated with some kind of
consciousness, he stressed that the purpose of conscious-
ness was to present its possessor with clear as opposed to
obscure ideas; this notion was built as a foundation stone
into the rational psychology of Wolff (1734/1972), who
anticipated late twentieth-century thought with his no-
tion that the purpose of mind was to give its possessor a
reliable representation (mental model?) of external real-
ity. The largely unconscious processes that presented the
subject with a clear picture of the external world at any
moment were subsumed under the general heading of
"apperception," which after Leibniz was elaborated on by
Wolff,. Kant, and Herbart before being made the central
point of Wundt's cognitive psychology; Wundt even
claimed that there was a special brain area devoted to the
apperception process. Fechner absorbed the notion of
extensive unconscious processing preceding the subjec-
tive experience of a sensation into his system, as noted in
the target article.
But both Gundlach and Gigerenzer also stress that
since mind and matter are two aspects of the same thing
we should not think that events in matter (brain-
substance) cause events in mind (sensations). Causation is
a descriptive word that should be reserved for discussing
events in matter only (e.g., physical stimulus> nervous
transmission brain event), and this is the topic of outer
psychophysics, which therefore has no need to refer to
sensations, according to Gundlach. Although this view is
maintained by others such as von Kries and Hornstein,
Gundlach claims that there is "no necessity to discuss
whether sensations are measurable," a more sweeping
claim that I shall return to below.
Stout claims that when I said that empirical questions
such as threshold measurement represent terra firma in
contrast to the slippery ground of questions about the
ontological status and scientific necessity of sensation
168 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
measurement, this is a false geology because the bedrock
of psychophysics is philosophy rather than threshold
measurement. But here Stout is somewhat at variance
with Gundlach, who claims that outer psychophysics,
concerned with such issues as threshold measurement, is
a valid scientific endeavour; and, in arguing that sensation
measurement might be futile, Gundlach seems to agree
with me that we are on slippery ground here. I believe
that the resolution of the apparent conflict is that I think of
sensation measurement as a marshland in a world
bounded by the goals of modern experimental psychol-
ogy, and of empirical measurements of thresholds, and so
on, as solid ground in this world. Stout is talking about a
different world, the world of Fechner's own metaphysics,
in which sensation magnitude was terra firma with re-
spect to the mental (a logarithmic function of the physical)
whereas the empirical measurement of the sensation
magnitude was slippery ground because of the number of
extraneous variables that made the measurement ofjnds,
and thence sensation, unreliable. Note, incidentally, that
Heidelberger claims that for Fechner the logarithmic law
seemed to contradict a kind of monism based on the view
that mind is an epiphenomenon of brain-substance; more-
over, other writers such as Mach and Hering placed the
locus of the logarithmic law at the periphery, not the
brain.
Treisman claims that in the target article I did not
acknowledge the difficulties of moving from arguments
about physics/physiology to arguments about the mental.
In particular, he suggests that we can take one argument
relating the physical to the mental, such as the argument
that equal stimulus differences Sx S2 and S3 S4 yield
equal mental experiences of apparent contrast, and then,
by a simple rephrasing of the relationship between S and
/, give what seems to be a different argument, namely,
that equal stimulus ratios SJS2 and S3/S4 yield equal
degrees of apparent contrast. However Treisman sub-
sumes this argument about the choice of the function
relating S and / under the more general argument about
being unable to make inferences concerning the transfor-
mation from / to E, E to S, or S to R given only a
knowledge of an empirical relationship between / and R. I
discussed this in the context of Shepard's (1981) views in
the target article (sect. 2.1), but I acknowledge that the
problem applies with equal force in the discussion of
whether contrast is judged in terms of sensation differ-
ences as opposed to sensation ratios. As von Kries (1882)
pointed out, we have to make arbitrary decisions in any
kind of measurement, but not only is there likely to be
less general agreement about the definition of a sone or a
dol than there is about the definition of a metre or a
second, there is also likely to be less general agreement
about contrasts between individual sensations because of
the arbitrary element involved in the specification of the
relationship between 1 and S.
As Marshall (1982) has shown, Fechner's philosophical
ambitions were spelled out in a monograph written when
Fechner was 22, and Heidelberger (1987) has stressed
that Fechner's monism must be seen in a broader context,
namely, his worries about whether a Laplacian determin-
ism can be applied to psychophysics (involving the men-
tal) or to studies of collective objects (such as samples of
the dimensions of paintings) in which measures of central
tendency (means, medians, modes) replace values of

variables as the numbers used in scientific propositions.
Cigerenzer stresses the importance of Heidelberger's
demonstration that the end result of Fechner's cogitations
was actually a form of indeterminism - the mental can be
indeterminate, and since the mental is another aspect of
the physical, the physical can also be indeterminate.
Heidelberger also showed that Fechner's theory of collec-
tive objects led indirectly to a redefinition of probability.
A probability value no longer was a measure of subjective
belief that a favourable event would occur, but could be
redefined as the limit to which the ratio of favourable to
unfavourable events would converge in an infinite sample
of events. Gigerenzer points to the "strange route" by
which Fechner's metaphysics of mind and matter contrib-
uted to the development of modern statistics. Gigerenzer
also points out that there is no name for Fechner's system.
Why not call his system "Fechnerism" (like "Darwinism"
or "Marxism") since it is unique, particularly in its specifi-
cation of the logarithmic relationship between matter and
mind?
It is precisely because this relationship is a dual-aspect
relationship rather than a causal one that Heidelberger
claims that "Fechner's Law" should not be called a law but
a "measurement formula" - as he notes, it is more akin to a
definition or an analytic statement. It is rooted in Fech-
ner's notion of measurement (and may owe something to
Fechner's study of Ohm, if Gigerenzer is right).
Heidelberger points out that whether Stevens's Law
should be called a law depends on whether we know the
measurement formula relating the sensation strength to
the subject's direct estimation of it. But it can be called a
"measurement formula" if it is derived from a fundamen-
tal formula such as Equation 2 or Equation 10. Fechner
(1887) argued that empirical tests were the way to decide
between Equation 2 and Equation 10 as fundamental
formulae, but he would, I think, have been as discon-
certed as I was when I read Luce's criticism of Fechner's
and my logic for getting from Equation 10 to Stevens's
Law. If I have understood Luce correctly, arriving at
Fechner's Law from Equation 2 by replacing AS/A/ by
dS/dJ is invalid for all Weber functions except A/// = k;
and in the argument from Equation 10 to Stevens's Law,
replacing AS/A/ by dS/d/ leads to a solution that is only
valid for the special case where the constant C in Equa-
tion 10 (which becomes the exponent of the power func-
tion) equals one.
If this is so, it helps resolve a problem I have run into in
teaching Fechner's Law to students with little mathe-
matical background. As described by Murray (1988, p.
182), in an argument similar to one used by William James
(1890/1950, vol. 1, p. 538), suppose for simplicity we take
the Weber fraction k as .5; then when a stimulus of
strength 8 is given a sensation strength of value 1, a just
noticeably different stimulus strength of 12 [i.e., 8 + (.5
X 8)] will give a sensation strength of 2, a stimulus of
strength 18 [i.e., 12 + (.5 X 12)] will have a sensation
strength of 3, and in general when we continue this
geometric series of stimulus strengths 8, 12, 18 . . . we
will have an arithmetic series of sensation strengths 1, 2,
3 . . . Since A times B is an (antilog) transform of log A
plus log B, sensation strength can be argued to be a
logarithmic function of stimulus strength given that We-
ber's Law is true.
The simplicity of the above argument hinges on the
Response/Murray: History of psychophysics
definition of the S-value of 2 as being one jnd above the
S-value of 1, that is, on the assumption that AS is constant.
If we try to derive a similar arithmetical argument from
Equation 10, which postulates that AS/S, rather than AS,
is a constant equalling C(A//7) we find that AS must
change in size as / increases and we can no longer assume
that S = 2 is one jnd (AS) above S = 1. Moreover Luce
says that we will only arrive by this route with a power
function relating S and / if C = 1. This explains why it is
easier to teach a class of students how to get from Weber's
Law to a logarithmic law than it is to teach them how to get
from Weber's Law to a power law.
The value of Fechner's contributions to the history of
psychophysics is still a matter for debate. Laming gives
reasons for thinking that psychophysicists have tended to
be trapped in a Fechnerian framework, particularly one
incorporating the notion of a fixed threshold; he notes the
virtues of signal detection theory, and of Quick's equa-
tion, for escaping the trap. On the other hand Link, both
in his commentary and in his recent book The wave theory
of difference and similarity (1992), has brought to our
attention the fact that Fechner's theory of discrimination
involved a rigorous model of the comparison process.
Fechner incorporates into his numerical treatment of data
not only the just noticeable difference (the difference in
magnitude between two stimuli that can be correctly
discriminated with probability of .75) but also an estimate
of the probable error. Link points out that the measure of
sensitivity resulting from the probable error was based on
the Gaussian theory of error and anticipated the form of
signal detection theory known as ideal observer theory; it
also anticipates Thurstone's theory of discrimination. I
would here draw the reader's attention to the first half of
Link's book, which traces the history of theories of dis-
crimination from Fechner to the present day and con-
siders latency, as well as accuracy, data. This literature
brings out forcibly the advantage for data interpretation
when experimenters provide a model of the mental pro-
cesses involved in comparing objects, as opposed to the
case where they passively retail discrimination data. In
the second half of his book, Link provides his own model
of comparison processes pertaining to simultaneously
presented sensory stimuli, but it differs from Fechner's
insofar as error is not seen as an explanation of discrimina-
tion difficulties; nor does it require that stimuli be or-
dered on a psychological continuum. Instead, the deci-
sion process is driven by differences being noted between
electrical signals that are the results of responses by
discrete receptive units (Link 1992, p. 199). In the light of
these recent contributions both by Laming and by Link,
Fechner's contributions will probably be seen by future
historians as signposts rather than destinations.
R2. Substantive questions about psychophysics
R2.1. The validity of Fechner's law. The logarithmic law
came under attack in Fechner's own time; not only did
Plateau and Brentano favour a power law under the belief
that equal ratios of physical intensities yielded equal
ratios of sensation magnitude, but there was also a general
feeling that the increment that would be added to a
sensation magnitude by a just detectable physical differ-
ence was itself proportional to the sensation magnitude in
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 169
Response/Murray. History of psychophysics
question. Hering was the standard bearer for this tradi-
tion, the pros and cons of which were discussed by
Delboeuf (1883) in a monograph whose long title included
the words "Hering against Fechner." In his commentary
Grusser indicates that Hering was more prescient than is
generally realised. As Grusser narrates, Hering's disillu-
sionment with Fechner not only stemmed from Hering's
experiments with weights (as described in sect. 1.3.4 of
the target article) but also from Hering's failure to obtain
an increasing linear relationship between apparent
brightness and log stimulus intensity; instead, Hering
obtained an S-shaped curve. Grusser not only points out
the significance of this curve in the development of
Hering's ideas about assimilation and dissimilation pro-
cesses, but he also points out the similarity between
Hering's curve and the Naka-Rushton function; in Figure
3 of the target article, a similar curve is derived from
Delboeuf s 1873 data by a transformation (it is not "fitted"
to Delboeuf's data). We are all indebted to Grusser for
pointing out Hering's contribution, but in endeavouring
to decide between a logarithmic law, a power law, and
"Hering's Law," as Grusser wishes to call it, we note that
it is at the extremes of the range of intensities that
Hering's Law yields predicted data that are most different
from the others. One of the lessons of Note 3 is that for
middle-of-the-range intensities the different laws can
yield almost equally good fits to data. But, as Teghtsoo-
nian emphasizes, the fit of the power function is predi-
cated on the notion that the subject adjusts Z3 so that S3/S2
= S2/Si, any three values of a variable forming a geomet-
ric series will be consistent with this notion as well as with
the notion that subjects adjust differences if logs of the
values are the basis of their judgments.
The argument that sensation increments are propor-
tional to sensation magnitude was central in Stevens's
development of the power law (Stevens 1957) and Olsson,
Harder & Baird stress the value Stevens placed on
Ekman's data of the 1950s. But they also stress that
Ekman's view as to whether sensation increments were
proportional to sensation magnitude depended on the
sensory dimension involved, with a constant AS more
likely to be found for brightness and pitch rather than for
taste. Moreover, if the subjective jnd were defined as a
standard deviation (a procedure reminiscent of Fechner's
use of variability in establishing the truth of Weber's Law;
see Fechner 1860/1966, pp. 85-93), the standard devia-
tion (an indirect measure of what we have called AS) was
shown by Ekman and Kiinnapas (1957) to vary with
sensation magnitude in an equation with three constants:
In the notation of Olsson et al. but with S referring to
sensation magnitude,
= cSm + b (18)
The so-called Ekman's Law, in the notation of the target
article, would be
AS = K'S (19)
where K' is a constant. Equation 19 is only equivalent to
Equation 18 if m = 1, b = 0, and K' bears a relationship to
S reflective of the relationship of AS to crs.
If Equation 18 is accurate, leaving a variety of ways in
which AS could vary with S, different exponents found
when power functions are fitted to sensory magnitudes
associated with different sensory dimensions may be
170 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
difficult to explain in terms of Weber's Law, a caveat
already noted by Olsson et al. These exponents may be
more related to the type of measurement involved in the
sensory dimension in question (see below).
The generalizability of the power function to certain
sensory dimensions has been questioned. For the case of
lifted weights, Ekman (1956), as noted by Olsson et al.,
found a linear relationship, with an additive constant,
between AS and S, but Ross notes that most of the studies
used by Stevens and Galanter (1957) appear not to have
taken account of the size-weight illusion. When weights
are lifted unseen, the function relating judged magnitude
to actual weight can be either linear (Anderson 1972;
Engen & Tulunay 1957) or logarithmic (Harvey & Camp-
bell 1963; Parducci 1963). Poulton (1989), who reviewed
this question, noted that whether linear or logarithmic
functions were obtained depended on the psychophysical
method: Interval or partition judgments tended to yield
linear functions; rating methods tended to yield log-
arithmic functions. As Ross hints, if we can eliminate
lifted weights (seen) and electric shock from the list of
stimulus dimensions associated with exponents greater
than one, we narrow the range of possible psychophysical
functions by excluding functions that are positively accel-
erated.
Brigner takes me to task for not having said enough
about evidence from the neurophysiological literature
related to sensory magnitude. He notes that Ratliff and
Hartline (1959) and Patton (1960) show that stimulus
intensity seems to be reflected in the logarithm of the
frequency of nervous impulses, a finding that had been
obtained earlier by Matthews (1931) for the case of the
muscle spindle. Werner and Mountcastle (1965), on the
other hand, showed that stimulus intensity might be
reflected in the frequency of nervous impulses raised to a
power, and Stevens (1970) collected as much evidence as
he could supporting a power law for neurelectric phe-
nomena. Then we have Lipetz (1971), however, who
reports data supporting the view that the Naka-Rushton
function (the tanh log function) is able to account for
linear, power, and logarithmic relationships between
stimulus strength and firing frequency (see also Griisser's
commentary). Brigner is right: The question of the appro-
priate psychophysical law applies to neurophysiological
data as well as to psychological data. But we must bear in
mind Handel's remark that by the time the sensory
nervous messages get to the brain level "there is no subset
of fibers that faithfully represents the firing pattern at the
receptor level."
The power law is in fact being buffeted from a number
of directions: As noted above, it is difficult to derive from
Weber's Law, although it is derivable from another route,
the assumption that equal sensation ratios correspond to
equal stimulus ratios (Stevens 1957; Yilmaz 1967); follow-
ing Ekman, the meaning of the exponents found when
power functions are fit to data is open to question; in the
case of lifted weights, there are reasons to question earlier
estimates of certain widely quoted exponents; and neur-
electric data have been fitted by logarithmic or Naka-
Rushton functions as well as by power functions. This
seems an appropriate juncture to stress an argument of
Poulton (1989), which I overlooked when writing the
target article. In his book Bias in quantifying judgments,
Poulton devotes a chapter to showing how biases in

responding can lead experimenters to mistakenly believe
that power functions are suitable descriptions of psycho-
physical functions.
First, we must describe what is meant by a "logarithmic
response bias." In magnitude estimation or category rat-
ing tasks, if subjects are given a large range of numbers as
possible responses, they will use 1, 2, 3 . . . n-digit
numbers about equally often, particularly when judg-
ments are being1 made about sensory dimensions that do
not have familiar units (e.g., brightness, loudness). Thus,
on any scale on which there is a step change in the range of
numbers available (e.g., 1, 10, 100), the effect will be to
make subjects give as responses numbers between 0 and
9, numbers between 10 and 99, and numbers between
100 and 999 about equally often, even though the range of
numbers comprised within each of these limits increases
in the order mentioned. Other circumstances in which
there is a logarithmic response bias include descending
judgments and cases in which instructions are given to
judge ratios; as Poulton says (p. 173), "these instructions
encourage the undergraduates to use for responses a scale
of numbers calibrated logarithmically instead of linearly."
Because this bias is so pervasive, Poulton recommends
using linear ascending judgments as the only sure proce-
dure for avoiding the logarithmic response bias.
The effect of this bias, according to Poulton, is to give
experimenters the impression that a power function is
operative, because when log estimated magnitude is
plotted against log stimulus intensity, a straight line often
results. This relationship, however, particularly in the
case when stimulus magnitudes cannot be judged in
familiar units, can be produced by the logarithmic re-
sponse bias or a range-determined bias. Poulton shows
how straight lines on log-log plots can be produced by
experimenters if they choose the right range of numbers
to be used as responses. Similarly, cross-modal matches
between two stimulus dimensions can be made consistent
by appropriate choices of the range of responses to be
used for each dimension. Since the size of the exponent
depends on the range of responses the experimenter
uses, exponents cannot be said to be unique to particular
sensory dimensions. As for the case of loudness, which
was historically the dimension for which Stevens claimed
that a logarithmic psychophysical function did not hold,
Poulton (1969) indicated that when range effects were
controlled, the discriminability of loudness for noise in-
creased linearly with log sound pressure level, whereas
for the loudness of tones, this psychophysical function for
various methods tended to be intermediate between
linear and logarithmic (Poulton 1989, Fig. 4.2). The
reader finishes Poulton's book with a strong scepticism
about the validity of a power law and a renewed feeling
that perhaps the logarithmic law has been rejected too
hastily; in Figure 1 of the target article the question mark
under the line heading from the logarithmic law might be
supplemented by "Poulton 1989."
Krueger notes that Delboeuf's Law is more suitable for
dealing with certain contrast effects than is a power law
(though see sect. R2.5 below) or a logarithmic law. I
would also note that Norwich (personal communication)
points out a resemblance between Delboeuf's Law,
which can be written
S = log[l + lie]
Response /Murray: History of psychophysics
and Norwich's own psychophysical equation
S = yl")
where k", y, and n are constants. This equation takes the
form of Fechner's Law when yln > 1 and Stevens's Law
when yln 1 (Norwich 1991, pp. 160-63).
R2.2. Signal detection theory. I am grateful to Laming for
arguing explicitly what was only implicit in the target
article, namely, that signal detection theory arose inde-
pendently of psychology in the contexts of engineering
and statistical decision making; Laming's Figure 1 would
suggest that I should have elaborated my Figure 1 to
shown an extra branch, a line reading from Rice, Shan-
non, and Fisher to the point labelled "signal detection
theory: Green & Swets 1966." Wagner also thought I
stretched things too far by relating signal detection theory
to inner psychophysics; the "noise" in the former theory is
a mathematical construct reflecting variability that is
"only peripherally" related to neural activity. These quali-
fications to my remarks in section 1.5 of the target article
are noted with appreciation.
However, I had not intended readers to believe I
thought signal detection theory was in any essential way
based on the assumption of normal distributions, and I
welcome Laming's demonstration, in his Figure 2, of the
good fit of a distribution that is not normal (in this case, it
is a noncentral x 2 distribution) to some data of Swets et al.
(1961). In fact, the generalizability of the signal detection
model to nonnormal distributions has been amply at-
tested to in the book by Egan (1975). I remain with my
claim, however, that we do not have enough data to
supplement the large theoretical edifice that has been
erected around signal detection theory in works such as
those by Green and Swets (1966), Egan (1975), and
Macmillan and Creelman (1991).
I put Laming's Sensory analysis in the neurelectric
column of Figure 1 because the book (and also the
multiple review, BBS 11(2) 1988) deals in detail with
probabilistic events that occur in the eye, ear, and other
sense organs. I am happy to acknowledge that its rightful
lineage is signal detection theory, however, and in fact
this is where I originally placed it in earlier drafts of the
target article.
R2.3. Measurement theory. When I wrote the target
article, I thought I was relatively isolated in worrying
about whether sensation can be measured, but section
1.5.3 elicited considerable commentary. In the article, I
mentioned von Kries as being at the centre of the move-
ment whose aim was to discredit the notion that sensa-
tions could be measured, but I would stress again that von
Kries was relatively late in the line of Fechner's detractors
on this issue: The neo-Kantians, for example, objected to
the idea that sensation could be isolated at all from the
apperception process; and in France certain mathemati-
cians had argued before von Kries that sensations were
not measurable in the way that lines were (Heidelberger
1993a). In this context the distinction should be made
between "extensive" measurement, where a value of a
magnitude on a numerical scale (say 6 units) can be
derived from adding together two shorter magnitudes
(say 3 units each), and "intensive" measurement, where
this is not possible. Length, for example, is an extensive
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 171
Response/Murray: History of psychophysics
magnitude: I can lay two 3-metre sticks end-to-end to
produce a 6-metre length. Temperature is an "intensive"
magnitude: I cannot add a beaker of water at 10C to
another at 10C and produce a new volume of water at
20C. Von Kries stressed that length was subject to the
concatenation criterion, as just outlined, and claimed that
sensation strength, being intensive, could not be mea-
sured given a concatenation criterion. Ramul (1960, p.
258) shows that Ploucquet in 1764 had also claimed that
psychological magnitudes, being intensive, could not be
concatenated.
In my target article, I gave a quotation from Mach
which seemed to support von Kries because it said "there
can be no question of an actual measurement of the
sensations; all that can be done is to characterize them
exactly and make an inventory of them by numerical
means." This quotation is from Mach's most widely read
book, and even Stevens (1975, p. 59) stressed the first
sentence, although he read the second sentence as imply-
ing that sensation could be categorized by means of
numbers. However, both Heidelberger and van Brakel
have hastened to correct me: By "measure" Mach meant
"measure according to the concatenation criterion" and
by "make an inventory" Mach meant "assign numbers to
different degrees of sensory magnitude in the way that
numbers can be assigned to different degrees of the
intensive magnitude of temperatures." Or, rephrasing it,
Mach was sceptical about whether sensation could be
measured on a ratio scale (whose attractiveness should not
be seen as limited to the fact that concatenation is an
analogue of the arithmetical operation of addition, since
other arithmetical operations can also give rise to mea-
sures concordant with concatenation, as both Luce and
van Brakel note).
I have two comments. First, Mach has a tendency to
use words idiosyncratically. Consider this passage from
his article entitled "Some questions of psycho-physics"
(Mach 1891)
Now I say, that if I see a tree with green leaves (A), with
a hard (B), gray (C) trunk, that A B C are elements of the
world. I say elements - and not sensations, also not
motions - because it is not my purpose at this place to
arrive at either a psychological or a physiological or a
physical theory, but to proceed descriptively. . . . If,
however, I close my eye (K), withdraw my feeling hand
(L),ABC . . . disappear. If I contemplate A B C in f/iis
dependence they are my sensations. This is but a
special point of view within the first.
According to my conception, therefore, the same A B
C is both element of the world (the "outer" world
namely) and element of feeling, (pp. 394-95)
This is a difficult passage to read at first because there is a
natural tendency to read Mach's definition of "sensation"
as referring to a memory image; but I think he is saying
that whether or not the green leaves of the tree are objects
or sensations depends on your point of view, and that one
can only tease apart a "sensation" of an object from the
object itself by the artificial experiment of closing one's
eyes. In the same paper, Mach objected to Fechner's
comparison of matter and mind as being "as inseparable as
the concave and convex sides of the same circle"; Mach
says this makes a duality where in fact unity exists, a
sophisticated form of monism. Mach's ideas were not
easily understood in his time, and, as Heidelberger and
172 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
van Brakel indicate, his sentences have to be understood
in their broad context.
However, my second point is more worrying: If Mach
said that sensations can only be categorized on an ordinal
scale, is there any hope that a psychophysical function will
have scientific validity? Temperature can certainly be
predicted from calculations about movements of mole-
cules; so it should be possible to predict sensation magni-
tude from a knowledge of stimulus intensity. The doubt
Mach raises is whether any inference about whether one
sensation magnitude is "twice" another is valid; and, of
course, as Wagner reminds us, this is what William James
had in mind when he said that one red is not a sum of
pinks. Similarly, if two liquids of different temperatures
are combined, we can predict the resulting temperature
at different intervals after the mixture is made, knowing
the laws of cooling; but how do we mix two sensations?
We can compare them, as we move a finger from a vessel
of hot water to a vessel of cold water; but these are "be-
fore" and "after" sensations, not a "mixture" of sensations.
Luce's contribution to a resolution of the dilemma is to
argue that the mistake psychophysicists make is to try to
measure just one dependent variable, sensation, when
only one independent variable, physical intensity, is
manipulated. He says that inferences made about the
effects of single independent variables are only successful
when the dependent variables are extensive measures of
physics such as mass, length, charge, and time, all of
which can be concatenated. Luce cites his work on sensa-
tion measurement when two or more independent vari-
ables are manipulated, an approach also espoused by
Anderson. Mach says that sensation magnitudes cannot
be concatenated; Luce writes that "unless a relevant
concatenation operation or some other rich internal struc-
ture can be found, the situation [the one-dimensional
measurement case] is simply too underdetermined to be
of much theoretical interest." If underdetermination, like
controversy, leads to overspeculation, William James
(1890/1950, vol. 1, p. 549) was vindicated in his worrying
about the "dreadful literature" of psychophysics.
One way in which the problem of sensation measure-
ment of a single sensory magnitude can be bypassed is by
studying the behaviour of subjects confronted by two
kinds of sensation, as in the cross-modal matching (CMM)
paradigm. Teghtsoonian in fact claims that CMM is the
answer to the riddle of how to measure sensation strength
when it is so hard to define it on a one-dimensional scale.
CMM avoids the "ancient questions" about ratios or
differences between sensations, questions that, as we saw
in the target article, have caused unusual difficulties. My
main problem with Teghtsoonians argument is that, as far
as he knows, "no one has ever suggested that the power
law is found for CMM only because the experimenter
induced a particular bias in his subject." Poulton (1989,
pp. 271-75), however, indicates how biasses operating in
Stevens's experiments could have induced subjects to
respond in such a way as to yield power functions for the
individual modalities investigated.
Hornstein, in suggesting that von Kries's argument was
unfairly ignored for over a century, says that my dismissal
of von Kries's critique reveals a "surprising presentism" in
my approach. First, I don't think I ever "dismissed" it; as I
wrote the target article, which was meant to be a "per-
spective," that is, a mental framework for encapsulating

the history of psychophysics in the reader's mind, I gave
von Kries a prominent place in the history of measure-
ment theory. Now, I may freely say that I think the
commentaries I have just reviewed have been extraordi-
narily helpful in improving that "perspective": We now
know, from Heidelberger and van Brakel, that von Kries
did not properly discuss the idea that sensation strength
could be measured on an ordinal or interval scale, and
restricted his idea of measurement to one permitting a
concatenation criterion. We have also learned from Luce
that von Kries's views fit into a particular place in the
context of the whole of science. Nevertheless it must be
acknowledged that there are still persons who object to
the idea of measuring sensations, either because they
believe the word "measure" has to be fitted in a Procrus-
tean way to the meaning of the word "sensation" (Horn-
stein likens this to the Red Queen's twisting of meaning)
or because sensations themselves are "ontologically du-
bious characters" that have no place in outer psycho-
physics (as Cundlach claims).
Second, I am becoming worried by the growing use of
"presentism" as a negative term meaning "biassed his-
tory. " It is impossible to write a history of a large topic
which is unbiassed by one's knowledge of the present
state of the topic, because presumably in a history of a
topic as large as psychophysics, I cannot simply convey
information about every paper without making some
evaluation of its usefulness from a present-day point of
view. There are indeed histories of other large topics in
psychology that rigorously avoid any kind of presentism.
Jackson's splendid history of ideas about depression down
the ages (Jackson 1986), for example, takes us through the
ancient and medieval periods with little reference to
modern theories, and this objectivity persists in his dis-
cussions of nineteenth- and twentieth-century theories of
depression. The end result is that the reader feels a strong
sense of the sheer weight of theorizing about depression
and yet also feels a sense of incompleteness (one, I hasten
to add, that is justified) by the lack of any feeling that we
are close to a simple explanation of depression. The
history of psychophysics also leaves one with a sense of
incompleteness. However, I think there is room for a
particular kind of presentism, one which takes a current
opinion and looks back through history for evidence or
theories that support it (as well as evidence against it).
The first half of Link's (1992) book is like this, and it leaves
the reader with a sense of progress rather than the lack of
it. But modern psychophysics, alas, is in such a state of
flux that instead of saying "this is the truth: Here is the
evidence that supports it," I have been obliged to say:
"These are the issues: Here is the evidence adduced on
these issues," leaving readers, of course, with the sense
that psychophysics is still too controversial for them to
base any strong reliance on any one of its theories.
However, if the word "psychophysics" is extended to
include any studies in which subjects make responses
concerning what they sense (perceive), then psycho-
physics is not such a disappointment as the target article
makes out. As Whittle writes, you can do psychophysics
without some prior commitment to an ideology about
sensation strength or a particular psychophysical law.
R2.4. Do we need "sensation strength" in a valid psycho-
physics? The first part of section 2.1 of the target article,
Response /Murray: History of psychophysics
on psychophysical methods, yielded few comments con-
cerning the classification of the methods shown in Tables
3 and 4. Teghtsoonian notes that I had not referred to
cross-modal matching, which would presumably fit in
Table 3 (nonverbal responses) with two stimuli (A on G, A
clearly discriminable). There would be a continuous
range of stimuli in another modality, B on a background
H, which have to be adjusted to match A. Borg stresses
that for purposes of generalizing beyond measures of
sensation strength to such measures as perceived exer-
tion, ratio scales or category ratings are useful, partic-
ularly when it comes to establishing a range from "lowest"
to "highest." Another classification of psychophysical
methods, which does include cross-modal matching, will
be found on pp. 39-40 of Poulton (1989). Here the
methods are classified into five broad headings: discrimi-
nating and matching judgments; interval or partition
judgments; category ratings; arithmetic judgments; and
cross-modal judgments. Choosing between classifications
will be a matter of personal taste; in the history of
psychology, a similar question about classifications was
extensively discussed in the nineteenth century, namely,
the question of the classification of the emotions. Many
writers from early times, including Aquinas, Descartes,
Hume, and Thomas Brown (1820/1830), had classified the
emotions in different ways; it was perhaps Bain
(1859/1880) who laid the fashion to rest when, in his The
emotions and the will, he summarised previous classifica-
tions and rejected them all as armchair exercises. Since
then, with the important exception of J. B. Watson (with
his three basic emotions of rage, fear, and love; Watson
1930, p. 138), classifying emotions has become a lost art.
Yet I feel that my armchair exercise in classifying the
psychophysical methods has the important spinoff of
forcing us to realize that all psychophysical judgments are
made against a sensory background and that the more
flexibility given to the subject, as in n-choice responses,
the more information the experimenter will obtain,
though possibly at the cost of giving the subject more
opportunity for bias.
On the other hand, the middle section of section 1.2
aroused more comment than any other passage in the
target article. Borg wishes I had elaborated on my remark
that "if we want a science in which assertions about
individual psychological experiences (e.g., "I don't like
stimulus X . . ,")are to be meaningfully related toothers
(". . . because stimulus X feels too intense"), it is difficult
to see how we could do without the expanded body of
information [involving the concept of sensation
strength]." Teghtsoonian points out that since the choice
of a reference continuum for describing sensation
strengths is entirely arbitrary, the relations that are ob-
tained among "sensation strengths" will depend on the
choice made. Whittle points out that hundreds of studies
of visual science are carried out with no mention of
"sensation strength" or "psychophysical laws"; he gives
convincing examples from his own work indicating that
"local" phenomena of sensation (such as brightness con-
stancy) are often not generalizable into comprehensive
assertions concerning other situations with ostensibly
similar characteristics. Lebedev, on the other hand, is
optimistic that we shall be able to relate assertions about
"sensation strength" to brain processes that show regular
cycles in time. Lockhead confirmed me in my suspicion
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 173
Response/Murray: History of psychophysics
that we should be more worried about "sensation dis-
tinctiveness" than about "sensation strength" when he
emphasized Brunswik's insistence (and, I add, the Gestalt
psychologists' insistence) on the difficulty of isolating the
effects of a single stimulus from those of the stimulus-in-
a-context.
I shall try to give some sense of closure to this general
worry about the usefulness of the concept of "sensation
strength" by arguing that we are here dealing with a
problem that is not so much about the neuropsychology of
perception as about the philosophy of science. It is, I
think, a mistake to believe that scientific propositions are
acceptable only if they are coherent and true; in a valid
scientific account of mental events, propositions must not
only be coherent with other scientific propositions, and of
course "true" or "veridical," but they must also be com-
municable. If we try to make scientific propositions about
how a person rides a bicycle (to use an old example), we
can either make assertions about macro-objects (the per-
son, the pedals, the wheels) or about micro-objects (the
nerve cells in the person's brain, the atoms in the pedals
and wheels). Both sets of propositions may be coherent
(e.g., within a Newtonian framework of physics) and
accurate, but the latter set may involve millions more
propositions than the former and therefore be less com-
municable because excessively cumbersome. The aim of
science is to present a representation of reality that can be
communicated from one scientist to another who has
roughly the same range of background knowledge; to this
extent science has an arbitrary component forced on it by
the need to be communicable.
Consider the problem of "sensation strength" in this
context. If I am talking at a level involving psychophysical
terms such as "preference" or "desire" it is quite appropri-
ate to use words indicative of a psychological magnitude of
"sensation strength." I can say "I like strong curry more
than mild because I enjoy the stronger sense of hotness"
or I can say "I desire to hear music at a loud volume
because I feel more caught up in the composer's art."
These are valid representations of facts about my psycho-
logical experiences which are coherent with other peo-
ple's representations of their own minds (even though
they may have different preferences) and are true in the
sense that I am not lying or misinformed or deluded.
When comparing neurelectric responses to stimuli of
different intensities, the scientist can communicate his
results about the neurelectric responses to others without
referring to sensation strength even if these responses, as
in the experiments of Lebedev, are measured in adult
waking subjects. If the scientist wishes to cross the gap
from the aspect of reality that consists only of things (like
stimuli and neurelectric responses) to the aspect of reality
that consists of mental experiences (like sensation) he is at
liberty to draw conclusions, based on apparent correla-
tions, concerning the possible relationships between
stimuli, neurelectric events, and "sensation strength" as
evaluated, say, on a rating scale or by my choice among
words describing different levels of sensation intensity.
These evaluations of "sensation strength" may be flawed
by biasses and may not be generalizable to persons other
than the subject giving the ratings, but within the context
of that representation of reality that focuses on the psy-
chological in relation to the physical, a proposition using
the word "sensation strength" can be true, coherent with
174 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
other evidence, and communicable (in the sense that
other people understand it and it is not cumbersomely
expressed). Once we accept communicability as a crite-
rion for the acceptance of propositions that purport to be
scientific, the way is open to an acceptance of "sensation
strength" as a term in a scientific psychology. Experimen-
tal problems associated with its measurement, however,
will lead to contentious discussion within the branch of
science that includes the term. Because "sensation
strength" is a referent in a representation of reality that
includes subjective experience, there will always be some
questions as to how to relate it to the representation of the
reality of physical objects, a fact that makes psychology
the most difficult of the sciences in the sense that, because
physics deals with objects, and sensations and thoughts
are not objects, psychology is difficult to integrate into
physics. More generally, any attempt to redescribe a
proposition about mental events such as "I do not like
stimulus X because it is too intense" in terms of physi-
ological events, even if true and coherent with other
physiological descriptions, would probably be too cum-
bersome to be useful in the representation of the reality of
the mental events. That everyday language is well suited
to encompass descriptions of sensation intensities is
clearly shown in the examples Borg gives of such descrip-
tions in the Bible and in Shakespeare.
On the question of teasing apart the effects of the
physical stimulus I, the neurelectric response , the
subjective sensation S and the final overt response R, I
note Lebedev's claim that S and E can be related to / in
terms of the time delay elapsing between the presenta-
tion of/ and the arrival of neural information related to / at
a point where the effect of/ is detectable in the electroen-
cephalographic record. Both Weiss and Anderson point
out that attempting to tease these variables apart for a
single stimulus overlooks the effect of the task and of the
stimulus-in-context; as Weiss says, an alternative view is
that "each stimulus domain and possibly each judgmental
task calls for independent investigation and validation of
the obtained scale. From this perspective, whether a
common law holds across continua is a question for the
data to decide." Anderson properly corrects me for mis-
stating his position (for which I apologize); the three main
issues for him do not involve the scaling of/ and S, which
are secondary factors in his information integration
theory, but the relationships between sensations prior to
integration, often nonconscious, the psychological law
describing the integration function, and the integrated
resultant, which is prototypically conscious. In a word,
there is a strong claim being made that the isolation of the
stages in the / S * R chain is an artificial exercise;
sensation and perception cannot be teased apart so easily,
a claim that we saw earlier can be traced back to Leibniz's
views on apperception. This issue must now be discussed
in more detail.
R2.5. Contrasts and contexts. It was in the middle of the
nineteenth century, when Helmholtz was expounding his
place theory of pitch sensation and his three-receptor
theory of colour sensation, that the evidence began to
accumulate that an account of sensory experience in
terms of a mosaic-like input from individual receptors was
incomplete. On colour contrast, for example, Helmholtz
and Hering agreed that the hue of a grey patch sur-

rounded by green differed from that of a patch of identical
grey surrounded by red; they disagreed on the explana-
tion, however, with Helmholtz favouring an account
based on perceptual learning and Hering favouring an
account in terms of neural interactions at a peripheral
level. Mach (1866) showed that a new sensation (a "Mach
band") could arise when a piece of paper divided into
equal black and white areas, but with a border showing a
sharp discontinuity in its contour, was rotated. The paper
looked mainly grey except for a darker strip or band at the
locus of the discontinuity. Some time later, von Ehrenfels
(1890) pointed out that the entity known as a "tune" is
based on relationships between adjacent items indepen-
dent of the individual frequencies of the tones. Evidence
like this persuaded the Gestalt psychologists of the early
twentieth century to stress that the "clear" percepts of
everyday experience are the end results of brain pro-
cesses that act automatically (probably innately) when-
ever the brain receives sensory information, in organizing
that information so that it becomes ecologically useful to
the percipient. In fact, almost the last sentence of Koh-
ler's (1969) last book summarizes the Gestalt question as
succinctly as any other: He asked "why do brain processes
tend to produce perceptual organizations of remarkable
clearness of structure?" This, of course, is almost identical
with Leibniz's question about clear and obscure ideas: In
this long-range historical perspective, Fechner's focus on
individual sensations can almost be seen as a diversion or
digression.
The importance of global factors in perception is
stressed by several commentators including Weiss and
Anderson as noted above; Gigerenzer and Lockhead,
both of whom support Brunswik's attempt to redefine the
perceptual process in terms of correlations between stim-
ulus events and mental events; Gregson, who claims that
psychophysical equations that might be valid for a single
stimulus-on-a-background, cannot simply be somehow
"put together" to yield equations valid for more than one
stimulus on the same background; and Handel, who
stresses how neurelectric events at the receptor level
cannot be simply tracked through the nervous system in
order to predict a neurelectric event at the brain level.
However, there seem to be two camps here: The above
commentators tend to be sceptical of any attempt to base a
theory of holistic perception on some kind of elaboration
of psychophysical laws applied to relationships between /
and R found for single stimuli; on the other hand,
Krueger feels that holistic effects such as contrast effects
should be avoided as far as possible in making determina-
tions of psychophysical laws; and Olsson et al. stress how
estimates of the size ofjnds can be different because of the
variability induced by other psychological factors. Which
view is correct?
Stevens himself, I think, would agree with Krueger
that there is a true psychophysical law relating the inten-
sity of a single stimulus (on a background) to the strength
of the associated sensation. For example, in order to
evaluate the effect of a change of overall illumination on a
single grey stimulus against a white background, Stevens
(1961) first took the stimulus on a background Bl and
showed that a power law described the relationship be-
tween the lightness of the stimulus and the magnitude
rating. But then he did the same again for the stimulus on
new backgrounds (B2, B3 . . . ). In each case a power
Response/Murray: History of psychophysics
function was obtained but the exponents changed with
the background. Then he took the same stimulus on a
fixed background Bl and changed the overall illumina-
tion: The effect of the overall illumination was to change
both the apparent lightness of the stimulus and the
apparent lightness of the background. He found that the
phenomenal appearance of the stimulus was predicted
from the earlier power functions relating / to R, given
each value of background. In particular, he found that
with a dark grey, the grey looked blacker as overall
illumination increased.
Since psychophysical power functions hold for each
stimulus on a given background, and these could be used
to predict appearances of the stimulus when both stim-
ulus and background changed because of the effect of
overall illumination, Stevens would claim that one can
take contrast effects into account and still make use of the
psychophysical laws applying to single stimuli. The point,
however, is that the exponent of the law in each case is a
function of the contrast between the stimulus and the
background, a point elaborated by du Buf (1987).
When is a stimulus ever judged in isolation, then? A
point of light in an otherwise pitch-dark room? A Ganz-
feld? I think many people would claim that experimental
data obtained under such circumstances are artificial in
that these circumstances rarely correspond with real-life
situations. Psychophysical experiments, in a word, are
claimed to lack ecological validity. The same claim has
been made for the experiments on letter or word identi-
fication using tachistoscopic presentation that seem to
support the notion of iconic memory (Haber 1983); we
never run into this situation in ordinary life (how often do
we read newspapers by lightning flashes?). And the claim
has also been made for experiments on memory using
nonsense syllables or isolated words (Bartlett 1932;
Neisser 1982).
The solution of rejecting all experiments involving
isolated stimuli as ecologically invalid is in my opinion too
harsh. Instead we need a comprehensive answer in which
variables that are undoubtedly influential in determining
the perception of isolated stimuli can be argued to be less
important in determining the perception of multistimulus
displays for a reason. In the case of perception, explana-
tions like this are in the offing, I think, if we consider such
work on contrast as that by Whittle (1992) and others. For
example, there is the phenomenon known as the "crisp-
ening effect," which is seen for isolated stimuli but is not
always apparent if the display is made more complicated.
If a circle of light is set up in a surround and the lumi-
nances of either are changed, there is a steep change in
the phenomenal brightness of the circle when its lumi-
nance passes through that of the surround. This crispen-
ing effect (Takasaki 1966) is possibly related to the fact that
sensation intensity at borders may not necessarily in-
crease with the logarithm of/ but with log A/, a finding I
should have mentioned in section 2.2 of the target article.
Whittle, however, has shown that the crispening effect
can be abolished by surrounding the circle with a thin
outline or by introducing hue differences between circles
and background. In his commentary, Whittle notes that
lightness constancy can be obtained if luminance ratios
are held constant in some situations but not in others; this
does not mean that the generalization has failed; it points
instead to the need for a more general interpretation in
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 175
Response/Murray: History of psychophysics
which the way subjects take account of overall illumina-
tion is a new variable. The study of "sensation" cannot be
isolated from the study of "contrast"; nor can the study of
"contrast" be isolated from global perception. Whittle
(1992), Gilchrist (1980), Hoffman and Dodwell (1985),
and other visual scientists are making substantial progress
in providing theories of global perception which explain
why phenomena obtained in experimental "artificial"
studies of stimulus isolation are not always apparent in the
viewing of everyday scenes.
I cannot resist adding that similar progress is being
made in the area of cognition: In artificial tasks such as
tachistoscopic word identification or word generation
(e.g., "name as many words as you can whose third letter
is R") it has been found that word frequency (the fre-
quency of words in large samples of text) is a major
determinant of performance (Murray 1975); in the real-
life production of sentences, however, it is the meaning of
the words, not their frequency, which determines what
words are produced. By banishing meaning from experi-
mental tasks like word identification or word generation,
we are forcing subjects to make use of a variable, word
frequency, whose effects are minor in real-life language
production. Here we are explaining why experiments
give results different from those in real life and, as with
the case of visual contrast, learning something in the
bargain. Psychophysical "laws" can be generated from
data relating / to R, but their usefulness in predicting
everyday behaviour is limited. Earlier I said I liked
"music" at high volume because I can better appreciate
the composer's art, but in real life this only applies to
music I like, such as that of Wagner or Beethoven, and not
to music I dislike, such as rock music. I suspect, but the
argument will need more elaboration, that quantitative
relationships between physical variables and psychologi-
cal variables are not too difficult tofindin situations where
stimuli are simple and isolated, but these quantitative
relationships cannot be extrapolated to many real-life
situations because meaning and other qualitative vari-
ables override quantitative variables such as stimulus
intensity in determining psychological judgments (a point
also hinted at by Marks). In the case of vision, variables
analogous to meaning in the cognitive realm would in-
clude knowledge of the object characteristics of the stim-
ulus, knowledge of the conditions of overall illumination
(including shadows), and Gestalt factors such as holistic
interpretations of individual parts of a complex stimulus.
R2.6. Use of response times in psychophysics. Three
commentators independently suggested that instead of
ratings, summated jnds, or other measures of sensation
strength, response times might be more valuable. Pe-
trusic reports some data by Cattell which indicated how
the response time to judge whether two lightnesses differ
decreases the greater the difference between the light-
nesses, with the relationship reflecting the compressive
nature of the psychophysical function for lightness. Pe-
trusic also tells us of Miinsterberg's little-known work on
response times and of his own extension of Miinsterberg's
paradigm. Lebedev says that "equal perceptual delays
correspond to equal sensations." (Interestingly, both Pe-
trusic and Lebedev feel that their data based on estimates
of time parameters are consistent with psychophysical
power functions.) Luce advocates the use of response
176 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
times as dependent variables because models of response
times are more process orientated; these models have
been worked out in impressive detail and are summarized
in Luce (1986), Townsend and Ashby (1983), and in
various chapters in Welford (1980). More generally,
Gregson points out that "a system with innumerable
internal elements does not have infinite dimensionality
and that nonlinear dissipative systems are in real time,
mathematically and empirically."
My immediate instinct is nevertheless to worry about
the sheer complexity of the task of trying to interpret
reaction times. Elsewhere I have expressed concern
about the difficulty of falsifying propositions in which the
time to yield a given response is said to be indirectly
influenced by a stimulus event (Murray 1987, pp. 75-76).
Here I shall stress the related problem that a minor
change in experimental manipulation can often alter the
timing of responses in such a way that, if one does
experiment X, one can never be sure the results will
generalize to experiments that differ slightly from X. We
can give three examples.
First, Luce (1986, p. 445) states that when two stimuli
(which may consist of a string of items such as letters) are
presented successively, with the stimuli being either the
same or different, the mean response time to give a
"same" response is faster on average than the mean time
to give a "different" response; I agree with Luce when he
finds this surprising because "only one difference needs to
be noted to warrant a 'different' response whereas all
elements must be identical to warrant a 'same' response"
(Luce 1986, p. 466). However in Figure 2 of Petrusic's
commentary, reporting response times when Munster-
berg presented two lines simultaneously, "same" judg-
ments were slower than "different" judgments for every
line length (and ratio of line lengths) examined.
Second, in the task of Sternberg (1966), a short list of
digits (or other stimuli) is presented succeeded by a single
digit and the subject must say whether the single digit
(the "probe") succeeding the list is the same as one that
was in the original list (is "old") or whether it is different (is
"new"). At least fourteen models have been presented to
account for the often replicated finding that the response
time to correctly say "same" ("old") increases linearly with
the number of digits in the original list (McNicol &
Stewart 1980, p. 256). Some of these models assume no
serial position effects (e.g., Sternberg's original serial
scanning model, 1966); some assume there are serial
position effects, so that items occurring late in the list are
responded to more quickly than earlier items - the linear
list-length effect can still be predicted (McElree & Do-
sher 1989; Murray 1986). But the key variable determin-
ing whether or not there will be serial position effects is
the delay of the probe. If it occurs very shortly after the
list, there are marked serial position effects; if it is delayed
over a second or so, serial position effects are much less
noticeable (Forrin & Cunningham 1973).
Third, a pattern that can emerge for response times
obtained when one kind of decision is required can
change if the decision is made more difficult. For exam-
ple, Chocholle (1940) showed that mean response time to
a tone decreased if the intensity of the single auditory
tone were increased; indeed, those who are interested in
changes in variability of psychophysical responses as a
function of stimulus intensity will be interested to note

the analysis done by Luce (1986, p. 446) of Chocholle's
data: As stimulus intensity increased, and therefore mean
response time decreased, the standard deviation of the
response time also decreased. But if we make the decision
harder by presenting one of two tones (of equal intensity),
each tone requiring a different response, and vary the
intensity of the two tones, it is found that response times
change in a U-shaped pattern as stimulus intensity in-
creases (van der Molen & Keuss 1979).
Other problems with response times include the scep-
ticism that can be engendered when we read the worries
by Luce (1986, p. 421) about our lack of understanding of
Hick's (1952) Law concerning the increase in choice
response time as the number of alternative stimuli in-
creases; or when we read his criticisms of the previously
acclaimed suggestion by Sternberg (1969) that analysis of
variance be used to estimate whether processes take place
at simultaneous or at successive stages (Luce 1986, p.
481); or when we read his criticisms of Donders's subtrac-
tion methods (Luce 1986, p. 213). I do not deny that the
study of response times might add rigour to our under-
standing of how sensations are processed; response times
are probably more reliable as data than are judgments
such as magnitude estimates; and one of the merits of the
recent wave theory of difference and similarity offered by
Link (1992) is his unification of judgmental and response
time measures within a single theoretical framework. But
response time data have to be approached with caution
because response times are so sensitive to any change in
experimental procedure.
On the other hand I think it is essential to note the
similarity of the ideas expressed by Lebedev to those
recently expressed by Newell in his Unified theories of
cognition (1990; see also multiple book review, BBS 15(3)
1992). Both authors indicate that different neurological
processes (reflecting different levels of cognitive activity)
tend to operate in time-quanta of approximately .01, . 1 , 1
or more seconds. That time units, rather than internal
scalable magnitudes, are the appropriate numbers to
include in cognitive science is an argument that is given
considerable credibility by the work of these writers.
R3. Concluding remarks
When I had finished the main part of the target article, I
had to write a concluding paragraph and found myself in
conflict between my desire to say that psychophysics had
made progress since Fechner and my desire to say, more
reluctantly, that in writing a perspective which took
Fechner as its starting point, I was forced to admit that I
frequently ran into stumbling blocks that made me less
positive in my evaluation of how useful psychophysics has
been. The latter sentiment won out (at the last minute)
and the final paragraph of my target article mentioned two
stumbling blocks in particular, the problem of whether
sensation strength can be measured and the problem of
whether we use sensation differences or ratios in judging
contrasts. These issues have also been discussed by a
number of commentators and several focused on the
matter of defining the units of sensation strength and
physical strength. It is worth summarizing these.
Borg notes that in a single experiment two subjects may
use different units of sensation strength, so that strictly
Response/Murray: History of psychophysics
speaking one cannot represent the results of the two
subjects on the same graph. Krueger notes that in select-
ing an anchoring point on a Fechnerian sensation scale,
the point named 70 is really an experimenter's "criterion"
rather than an observer's "threshold," while the choice of
units on the physical scale has been neglected in many
accounts of psychophysics even though it determines the
value of the exponent in any power function. Teghtsoo-
nian is so impressed with the arbitrary elements involved
in scaling sensation that he says the goal of measuring
sensation strength is a "will-o'-the-wisp," which my dic-
tionary defines as a light which misleads travellers, often
seen over marshy places (!). Treisman notes that once you
have defined the function relating S to I, whether or not
sensation differences or ratios are judged by subjects falls
out of the definition. Van Brakel points out that arbitrari-
ness applies not only to sensation scales, but also to
physical scales such as length; as he writes, "although,
allegedly, the only arbitrary choice for a ratio scale is its
unit, . . . the same unit can 'support' two different
scales, which are theoretically equally adequate." Wag-
ner claims that, just as instructions can determine how
perceived size seems to vary with physical size, so in-
structions can determine how "sensation strength" seems
to vary with "stimulus strength." Stout and Wagner note
that William James rejected Fechner's psychophysics on
grounds related to the difficulty of measuring sensations.
In the light of these comments, so-called measure-
ments of sensation strength seem to be among the less
valuable accomplishments in the history of psychophysics
so far. Another group of commentators, however, claims
that the exercise of measuring sensation strength might
also be of limited usefulness in the overall perspective of
studies of sensation and perception. Anderson claims that
Fechner did not properly appreciate the importance of
the way the nervous system integrates information from
multisensory displays even though, as I noted, previous
eighteenth- and nineteenth-century German work on
apperception can be seen to anticipate Anderson's views
on unconscious processing. Brigner finds Fechner's work
only incidental to the development of global theories of
inner psychophysics - the brain itself seems to provide a
context for some sensory experiences such as phantom
limb pain (and, I would add, dreams). Gregson (1988) has
offered a new perspective altogether for viewing psycho-
physics, one in which the dynamics of large networks
depend very much on local connections and not on the
sheer size of the networks; when the behaviour of such
networks is very simple, it yields behaviour almost look-
ing like traditional psychophysics, but it is in this context
that the psychophysical laws of Plateau and Fechner have
limited value. In his commentary, Gregson criticizes
these for not being "dissipative." Krueger.also notes how
extreme sensitivity to initial conditions can give rise, by
deterministic processes, to unpredictable and complex
behaviour. Handel shows neatly how the nervous system
is too complex for a simple "neurelectric law" that is
parallel to a simple "psychophysical law"; stimulus dura-
tion, for an example, is a variable ignored in psychophysi-
cal laws which actually determines the behaviour of
neurons at higher levels in the frog's nervous system.
Laming argues that Fechner "directed attention to some
inappropriate questions (to which answers are never
found); and his successors have become fixated on those
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 177
References/Murray: History of psychophysics
particular questions, unable to turn away, even when
experimental evidence so dictates." Laming quotes the
literature on fixed thresholds.
Lockhead claims that there are two literatures on
psychophysics, one which sees the role of the stimulus
environment as a nuisance contaminating the measure-
ment of the sensation corresponding to that stimulus (a
view that Brunswik called "structuralism"); and another,
Brunswik's "contextual psychophysics" as Lockhead calls
it, which would presumably include adaptation level
theory and Gestalt psychology as well as Brunswik's own
theory. In this second literature, Lockhead says that
"Fechner-like laws cannot be correct beyond being points
in a complex function where performance depends on
many factors which sometimes interact in complex ways."
Presumably Krueger's claim that "in order to reveal the
true underlying psychophysical function, it is necessary
to pare away bias and context effects not only at the
cognitive level, but also at the sensory or neural level" is a
dogma of Lockhead's first kind of literature, not the
second. Weiss, like Luce and Anderson, sees a psycho-
physics based on measurements of unisensory stimuli to
be a sort of "special case" which can fall out of a more
general system of functional measurement - "sensation
scale values fall out as a byproduct of successful model-
ling. A characteristic of functional measurement analysis
is that it calls for complex judgments, in which two or
more stimuli are integrated by the observer, presumably
in accord with the hypothesized model . . . Context ef-
fects . . . enrich the model . . . rather than [obscure]
pure estimates of sensation as the earlier tradition held."
Finally, Whittle sees Fechnerian theory as having pro-
vided methods and questions which once "seemed to be
substantive generalisations worth fighting over, [but] are
now seen to be local questions to be answered . . . en
route to something of a more current interest."
In the above, I have deliberately used quotations from
the commentators because they clearly indicate that
many psychophysicists now writing think of Fechner's
work as limited in scope. In the last 15 years or so, there
has been a revival of Gestalt-type thinking among stu-
dents of perception, and psychophysicists now - using
the word in Whittle's broad sense - wish to delve into
neuropsychology and bring in cognitive psychology in a
way that is foreign to traditional psychophysics. As Whit-
tle says, you can know a great deal about visual science
without worrying too much about whether Fechner's
questions are answered by that science. At the same time,
caution must be applied when drawing conclusions that
"holistic" or "global" properties dominate the perceptual
processing; Kimchi (1992) distinguishes between "global"
properties of a stimulus, defined by their position in the
hierarchical structure of the stimulus, and "holistic" prop-
erties, defined as a function of interrelations between the
component parts of the stimulus. Her review is a good
starting point for a more careful examination of the Ge-
stalt revival. Hoffman and Dodwell (1985) show how
global Gestalt laws can be derived from a "geometric
psychology" that postulates how the brain processes par-
ticular local responses.
When I submitted the manuscript of the target article,
one referee said that it might be useful as a teaching
device. The more I think about this, the more I am
178 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
content to see it as that, a device for acquainting young
persons who wish to do psychophysics with a framework
for thinking about this topic. It is clearer to me now,
however, than it was when I first wrote the article that it is
a framework limited to attempts to answer Fechner's
questions in Fechner's terms. It was he who stressed
thresholds, just noticeable differences, sensation differ-
ences and ratios, and the difference between inner and
outer psychophysics. The commentators have made it
quite clear that in the future psychophysics may see
Fechner's psychophysics as one (historically important)
aspect of a more comprehensive psychophysics in which
context effects are central, biasses to respond are properly
integrated into the topic, and neurelectric events affect-
ing cognition are not thought of as simple monotonic
transformations of neurelectric events at the periphery. A
psychophysicist of the future who wished to write his own
"perspective" would probably see my perspective as
Gregson sees it, the middle of a fish whose head (in the
past) lay in eighteenth-century theories of apperception
and whose tail (in the future) will incorporate the
twentieth-century writings of Helson, Brunswik, the Ge-
stalt writers, and Anderson, as well as new developments
in neurophysiology and in mathematical modelling.
References
Letters a and r appearing before authors' initials refer to target article and
response respectively.
Anderson, N. H. (1970) Functional measurement and psychophysical
judgment. Psychological Review 77:153-70. [aDJM, DJW]
(1972) Cross-talk validation of functional measurement. Perception and
Psychophysics 12:389-95. [rDJM]
(1975) On the role of context effects in psychophysical judgment.
Psychological Review 6:462-82. [DJW]
(1981) Foundations of information integration theory. Academic
Press. [RDL]
(1982) Methods of information integration theory. Academic Press. (NHA,
RDL, DJW]
(1990) Integration psychophysics. In: Psychophysical explorations of mental
structures, ed. H.-G. Geissler. Hogrefe & Huber. [NHA]
(1992) Integration psychophysics and cognition. In: Psychophysical
approaches to cognition, ed. D. Algom. Elsevier. [NHA]
Atkinson, R. C. (1963) A variable sensitivity theory of signal detection.
Psychological Review 70:91-106. [DL]
Atkinson, W. H. (1982) A general equation for sensory magnitude. Perception
and Pstjchophysics 31:26-40. [aDJM]
Bain, A. (1859/1880) The emotions and the will, 3rd ed. Longmans,
Green. [rDJM]
Baird, J. C. (1970) A cognitive theory of psychophysics: II. Fechner's law and
Stevens'law. Scandinavian Journal of Psychology 11:89-102. [MO]
Baird, J. C. & Noma, E. (1975) Psychophysical study of numbers. 1.
Generation of numerical responses. Psychological Research 37:281-
97. [aDJM]
(1978) Fundamentals of scaling and psychophysics. Wiley. [MO]
Baird, J. C. & Wagner, M. (1991) Transformation theory of size judgment.
Journal of Experimental Psychology: Human Perception b Performance
17:852-64. [MW]
Baker, K. E. & Dudek, F. J. (1955) Weight scales from ratio judgment? nd
comparisons of existent weight scales. Journal of Experimental Psychology
50:291-308. [HER]
Balmer, H. (1977) Albrecht von Holler. Haupt. [O-JG]
Bartlett, F. C. (1932) Remembering: A study in experimental and social
psychology. Cambridge University Press (Cambridge). [rDJM]
Bartley, H. S. (1970) The homeostatic and comfort perceptual systems.
Journal of Psychology 75:157-62. [GB1
Baxt, N. (1871/1982) On the time necessary for a visual impression to come
into consciousness. Psychological Research 44:1-12. (Translated by D. J.
Murray.) [aDJM]
Beck, J. (1972) Surface color perception. Cornell University Press. [aDJM]

Bell, E. T. (1939) Men of mathematics. Gollancz. [RAMG]
Berkeley, C. (1709) An essay towards a new theory of vision. Dublin: Pepyat.
(Reprinted in [1948] G. Berkeley's works, vol. 1, ed. A. A. Luce & T. E.
Jessop. Nelson.) [O-JC]
Bernoulli (1837) (See Fechner, 1860). [GB]
Birch, T. (1756/1968) The history of the Royal Society of London, vol. 3.
Jonson. [O-JG]
Bimbaum, M. H. (1974) Using contextual effects to derive psychophysical
laws. Perception and Psychophysics 15:89-96. [DJW]
(1980) Comparison of two theories of "ratio" and "difference" judgments.
Journal of Experimental Psychology: Ccneral 109:304-19. [aDJM]
(1989) Should contextual effects in human judgment be avoided?
Contemporary Psychology 37:21-23. [GG]
Bimbaum, M. H. & Veil, C. T. (1974) Scale convergence as a criterion for
reseating: Information integration with difference, ratio and averaging
tasks. Perception and Psychophysics 15:7-15. [aDJM]
Bjfirkman, M. (1958) Some relationships between psychophysical parameters.
Reports from the Psychological Laboratory, University of Stockholm, No.
65. [MO1
(1960) Variability data and direct quantitative judgment for scaling
subjective magnitude. Reports from the Psychological Laboratory,
University of Stockholm, No. 78. [MO]
Bjorkman, M. & Ekman, G. (1962) Exjjerimentalpsykologiska metoder
[Methods in experimental psychology], Almqvist & Wiksell. [GB]
Blackwcll, H. R. (1963) Neural theories of simple visual discriminations.
Journal of the Optical Society of America 53:129-60. [DL]
Blackwell, H. R., Pritchard, B. S. & Ohmart, T. G. (1954) Automatic
apparatus for stimulus presentation and recording in visual threshold
experiments. Journal of the Optical Society of America 44:322-66. [DL]
Bonne, C. (1989) How to bridge the gap between magnitude estimation and
reaction time. In: Fechner Day 90. Proceedings of the Sixth Annual
Meeting of the International Society of Psychophysics, ed. F. Muller.
Institute fur Psychologie Universitfit Wuerzburg. [ANL]
Borg, G. (1961) Interindividual scaling and perception of muscular force.
Kungliga Fysiografiska sallskapets i Lund forhandlingar 31:117-
25. [GB]
(1962) Physical performance and perceived exertion. Studia Psychologica et
Paedogogica, Series altera, Investigationes XI. Gleerup, Lund. [GB]
(1978) Subjective effort in relation to physical performance and working
capacity. In: Psychology: From research to practice, ed. H. L. Pick et al.
Plenum. [GB]
(1982) A category scale with ratio properties for intermodal and
interindividual comparisons. In: Psychophysical judgment and the process
of perception, ed. H. G. Geissler & P. Petzold. VEB Deutscher Verlag
der Wissenschaften. [GB]
(1990) A general model for interindividual comparison. In: Recent trends in
theoretical psychology, vol. 2, ed. W. J. Baker, M. E. Hyland, R. van
Hezewijk & S. Terwee. Springer-Verlag. [GB]
Borg, C., Burg, M. van den, Hassmen, P., Kaijser, L. & Tanaka, S. (1987)
Relationships between perceived exertion, HR and HLa in cycling,
running and walking. Scandinavian Journal of Sports Sciences 9:69-
77. [CB]
Borg, C., Diamant, H., Strom, L. & Zotterman, V. (1967) The relation
between neural and perceptual intensity: A comparative study on the
neural and psychophysical response to taste stimuli. Journal of Physiology
192:13-20. [GB]
Boring, E. G. (1917) A chart of the psychometric function. American Journal
of Psychology 28:465-70. [aDJM]
(1942) Sensation and perception in the history of experimental psychology.
Appleton-Century-Crofts. [rDJM]
(1929) A history of experimental psychology. Appleton-Century. [GG]
Bradley, J. (1971) Mach's philosophy of science. Athlone Press. [JVB]
Brcntano, F. (1874/1973) Psychology from an empirical standpoint.
(Translated by A. C. Rancurello, D. B. Terrell & L. L. McAlister).
Routledgc & Kegan Paul. [aDJM]
Brooks, G. P. & Aalto, S. K. (1981) The rise and fall of moral algebra: Frances
Hutcheson and the mathematization of psychology. Journal of the History
of the Behavioral Sciences 17:343-56. [rDJM]
Brown, J. L. & Mueller, C. C. (1965) Brightness discrimination and
brightness contrast. In: Vision and visual perception, ed. C. Graham.
Wiley. [aDJM]
Brown, T. (1820/1830) Lectures on the philosophy of the human mind.
William Tait. [rDJM]
Brunswik, E. (1934) Wahrnehmung und Cegenstandswelt: Grundlegung einer
Psychologie vom Cegenstand her. Deuticke. [GG]
(1956) Perception and the representative design of psychological
experiments. University of California Press. [GRL]
Re/erences/Murray: History of psychophysics
Buf, J. M. H. du (1987) Spatial characteristics of brightness and apparent-
contrast perception. Ph.D. dissertation, Technical University of
Eindhoven, Netherlands. [rDJM]
Buhler, W. K. (1987) Gauss, eine biographische Studie. Springer-
Verlag. [RAMG]
Biittner, U. & Grusser, O.-J. (1968) Quantitative Analyse der
Erregungssummation im receptiven Feld retinaler Neurone der Katze.
Kybernetik 4:81-94. [O-JG]
Campbell, N. R. (1920) Physics: The elements. Cambridge University Press.
(Reprinted in 1957 as Foundations of science: The philosophy of theory
and experiments. Dover.) [RDL]
Carnap, R. (1945) The two concepts of probability. Philosophy and
Phenomenological Research 5:513-32. [MH]
Cattell, J. M. (1893) On errors of observation. American Journal of Psychology
5:285-93. [aDJM]
(1902) The time of perception as a measure of differences in intensity.
Philosophische Studien 19:63-68. (Reprinted in [1947] James McKeen
Cattell: Man of Science, vol. 1. Science Press.) [WMP]
Chocholle, R. (1940) Variations des temps de reaction auditifs en function de
I'intensite a diverses frequences. L'Annie Psychologique 41:65-
124. [rDJM]
Churchland, P. M. (1990) Matter and consciousness. MIT Press. [GB]
Coombs, C. H. (1964) A theory of data. Wiley. [WMP]
Cornsweet, T. N. (1970) Visual perception. Academic Press. [aDJM]
Corso, J. F. (1963) A theoretico-historical review of the threshold concept.
Psychological Bulletin 60:356-70. [aDJM]
(1967) The experimental psychology of sensory behavior. Holt, Rinehart &
Winston. [aDJM]
Cross, D. V. & Rotkin, L. (1975) The relation between size and apparent
heaviness. Perception and Psychophysics 18:79-87. [HER]
Curry, R. E., Nagel, C. D. & Gai, E. G. (1977) Decision behavior with
changing signal strengths. Journal of Mathematical Psychology 15:46-
69. [aDJM]
Dale, A. I. (1991) A history of inverse probability. Springer-Verlag. [RAMG]
Daston, L. (1988) Classical probability in the Enlightenment. Princeton
University Press. [GG]
David, H. A. (1969) The method of paired comparisons. Griffin. [SL]
Dawson, W. E. & Miller, M. E. (1978) Inverse attribute functions and the
proposed modifications of the power law. Perception and Psychophysics
24:457-65. [aDJM]
de Lange, H. (1958) Research into the dynamic nature of the human fovea-
cortex systems with intermittent and modulated light. I. Attenuation
characteristics with white and colored lights. Journal of the Optical
Society of America 48:777-84. [aDJM]
Delboeuf, J. R. L. (1873) Etude psychophysique: Recherches theoriques et
experimentales sur la mesure des sensations et specialement des
sensations de lumiere et de fatigue. Mdmoires couronnis et autres
mfanoires publies par I'Academie Royale des Sciences, des Lettres, et des
Beaux-arts de Belgique, vol. 23. Brussels: Hayez. [aDJM]
(1883) Examen critique de la loi psychophysique: Sa base et sa signification:
Hering contre Fechner: Fechner contre ses adversaires. Librairie Germer
Bailliere et cie. [rDJM]
Delezenne, C. E. J. (1827) Sur les valeurs numeriques des notes de la
gamme. Recueil des travaux de la SocidtS des Sciences, de iAgriculture et
des Aris de Lille 1-56. [rDJM]
Diamond, A. L. (1953) Foveal simultaneous brightness contrast as a function
of inducing- and test-field luminances. Journal of Experimental
Psychology 45:304-14. [aDJM]
Drobisch, M. W. (1850/1972) Erste Grundlehren der mathematischen
Psychologie. E. J. Bonset. [rDJM]
(1861) Neue Ableit der Grundformeln von Fechner's Psychophysik. Berichte
uber die Verhandlungen der Koniglich Sachsischen Gesselschaft der
Wissenschaften zu Leipzig, 13. [RAMG]
Dusoir, A. E. (1975) Treatment of bias in detection and recognition models: A
review. Perception and Psychophysics 17:167-78. [aDJM]
(1983) Isobias curves in some detection tasks. Perception and Psychophysics
33:403-12. [aDJM]
Ebbinghaus, H. (1885/1964) Uber des Gedachtnis (Translated by H. A. Ruger
& C. E. Bussenius: Memory.) Dover. [rDJM]
Egan, J. P. (1958) Recognition memory and the operating characteristic
(Technical Note). Indiana University Hearing and Communication
Laboratory. [aDJMJ
(1975) Signal detection theory and ROC analysis. Academic Press. [rDJM]
Ehrenfels, C. von (1890) Uber Gestaltqualitaten. Vierteljahrschrift fur
wissenschaftliche Philosophic und Soziologie 14:249-92. [rDJM]
Ehrlich, P. (1982) Negative, infinite and hotter than infinite temperatures.
Synthese 50:233-77. [JVB]
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 179
References/Murray: History of psychophysics
Eisler, H. & Montgomery, H. (1974) On theoretical and realizable ideal
conditions in psychophysics: Magnitude and category scales and their
relation. Perception and Psychophysics 16:157-68. [MO]
Ekman, C. (1956) Discriminal sensitivity on the subjective continuum. Ada
Psychologica 12:233-43. [arDJM, MO]
(1959) Weber's law and related functions. Journal of Psychology 47:343--
52. [aDJM, MO]
Ekman, G. & Kiinnapas, T. M. (1957) Subjective dispersion and the Weber
fraction. Reports from the Psychological Laboratory, University of
Stockholm, No. 41. [rDJM, MO]
Ellis, B. (1967) Measurement. In: Encyclopedia of philosophy, vol. 5, ed. P.
Edwards. Macmillan. [MH]
(1968) Basic concepts of measurement. Cambridge University Press
(Cambridge). |JVB]
Engen, T. & Tulunay, U. (1957) Some sources of error in half-heaviness
judgments. Journal of Experimental Psychology 54:208-12. [rDJM]
Eriksen, C. W. (1966) Temporal luminance summation effects in backward
and forward masking. Perception and Psychophysics 1:87-92. [aDJM]
Falmagne, J.-C. (1985) Elements of psychophysical theory. Oxford University
Press. [aDJM]
(1986) Psychophysical measurement and theory. In: Handbook of perception
and human performance, vol. 1, ed. K. R. Boff, L. Kaufman & J. P.
Thomas. Wiley. [aDJM]
Fechner, G. T. (1848) Nanna oder iiber das Seelenleben der Pflanzen.
Leipzig: Voss. [GG]
(1851) Zend-Avesta: Oder iiber die Dinge des Himmels und des Jenseits.
Leipzig: Voss. [aDJM, GG, ANL]
(1858) Das psychische Mass. Zeitschrift fur Philosophic und philosophische
Kritik 32:1-24. [MH]
(1860/1964) Elemente der Psychophysik, vols. 1 and 2. Leipzig: Breitkopf &
HSrtel. (Reprinted: [1964] E. J. Bonset.) [aDJM, GB, GG, O-JG, MH,
SL]
(1860/1966) Elements of psychophysics, vol. 1, ed. E. G. Boring & D. H.
Howes. (Translated by H. Adler.) Holt, Rinehart & Winston. [arDJM,
SL, DAS]
(1866) Das Buchlein vom Leben nach dem Tode, 2nd ed. Leipzig: Voss.
(Originally published in 1836 using the pseudonym Dr. Mises.) [GG]
(1876) Vorschule der Aesthetik. Leipzig: Breitkopf & Hartel. [rDJM]
(1877) In Sachen der Psychophysik. Leipzig: Breitkopf & Hartel. [aDJM,
MH, LEK]
(1882) Revision der Hauptpuncte der Psychophysik. Leipzig: Breitkopf &
Hartel. [aDJM, MH]
(1887) Ueber die psychischen Massprincipien und das Weber'sche Gesetz.
Philosophische Studien 4:161-230. [arDJM, MH]
(1897) Kollektivmasslehre, ed. G. F. Lipps. Leipzig; W.
Engelmann. [aDJM, GG]
(1987a) Outline of a new principle of mathematical psychology (1851).
(Translated and edited by Eckart Scheerer.) Psychological Research
49:203-7. [aDJM]
(1987b) My own viewpoint on mental measurement (1887). (Translated and
edited by Eckart Scheerer.) Psychological Research 49:213-19. [aDJM]
Feng, A. S., Hall, J. C. & Gooler, D. M. (1990) Neural basis of sound pattern
recognition in anurans. Progress in Neurobiology 34:313-29. [SH]
Ferguson, A., Meyers, C. S., Bartlett, R. J., Banister, H., Bartlett, F. C. et
al. (1940) Quantitative estimates of sensory events. (Final report of the
committee appointed to consider and report upon the possibility of
quantitative estimates of sensory events.) British Association for the
Advancement of Science 2:331-49. [aDJM, GAH, JVB]
Fisher, R. A. (1922) On the mathematical foundations of theoretical statistics.
Philosophical Transactions of the Royal Society of London 222A:309-
68. [DL]
(1935) The design of experiments. Oliver & Boyd. [GG]
Fodor, J. A. (1968) Psychological explanation. Random House. [MH]
Foerster, M., Grind, W. van de & Grusser, O.-J. (1977) The response of cat
horizontal cells to flicker stimuli of different area, intensity and
frequency. Experimental Brain Research 29:367-85. [O-JG]
Foley, J. M. & Legge, G. E. (1981) Contrast detection and near-threshold
discrimination in human vision. Vision Research 21:1041-53. [DL]
Forbes, E. G. (1971) Tobias Mayer's opera inedita. The first translation of the
Lichtenberg edition of 1775. Macmillan. [O-JG]
(1972) The unpublished writings of Tobias Mayer. 3 vols. Vandenhoeck &
Ruprecht. [OJG]
(1980) Tobias Mayer (1723-1762): Pioneer of enlightened sciences in
Cermany. Vandenhoeck & Ruprecht. [O-JG]
Forrin, B. & Cunningham, K. (1973) Recognition time and serial position of
probed item in short-term memory. Journal of Experimental Psychology
99:272-79. [rDJM]
180 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
Foss, C. E., Nickerson, D. & Cranville, W. C. (1944) Analysis of the Oswald
color system. Journal of the Optical Society of America 34:361-
81. [aDJM]
Furman, G. G. (1965) Comparison of models for subtractive and shunting
lateral inhibition in receptor-neuron fields. Kybernetik 2:257-74. [O-JG]
Gamau, G. (1811) Erinnerungen an Lichtenbergs Vorlesungen Ober Erxlcbens
Anfangsgrunde der Saturlehre, vol. 2. Wien: Geistinger. [O-JG]
Garner, W. R. (1954) A technique and a scale for loudness measurement.
Journal of the Acoustical Society of America 26:73-88. [DJW]
(1962) Uncertainty and structure as psychological concepts.
Wiley. [aDJM]
(1974) The processing of information and structure. Erlbaum. [SH]
Gauss, C. F. (1809) Theoria motus corporum coelestium. Hamburg: Certhes &
Besser. [SL]
(1821/1880) Theoria combinationis observationum erroribus minimis
obnoxiae, pars prior. In: Carl Friedrich Causs Werke, vol. 4. KSniglichen
Gesellschaft der Wissenschaften zu Gottingen. [SL]
Geissler, H.-G. (1990) Foundations of quantized processing. In:
Psychophysical explorations of mental structures, ed. H.-G. Geissler.
Hogrefe & Huber. (ANL)
Geldard, F. A. (1972) The human senses, 2nd ed. Wiley. [HER]
Gescheider, G. A. (1976) Psychophysics: Method and theory.
Erlbaum. [aDJM, HER]
(1984) Psychophysics: Method, theory and application, 2nd ed.
Erlbaum. [SH, MO]
Gibson, J. J. (1979) The ecological approach to visual perception. Houghton
Mifflin. [SH]
Gigerenzer, G. & Murray, D. J. (1987) Cognition as intuitive statistics.
Erlbaum. [aDJM, GG]
Gilchrist, A. L. (1980) When does perceived brightness depend on perceived
spatial arrangement? Perception and Psychophysics 28:527-38. [arDJM]
Goldner, K. (1983) Mathematische Grundlagen der Systemanalyse, vol. 2.
Verlag Harri Deutsch. [RAMG]
Graham, N. (1977) Visual detection of aperiodic spatial stimuli by probability
summation among narrowband channels. Vision Research 17:637
52. [DL]
(1989) Visual pattern analyzers. Oxford University Press. [PW]
Green, D. M. & Swets, J. A. (1966) Signal detection theory and
psychophysics. Wiley. [arDJM, DL]
Gregson, R. A. M. (1988) Nonlinear psychophysical dynamics.
Erlbaum. [RAMG]
(1991) Dead men tell odd simple tales! Behavioral and Brain Sciences
14:187-88. [aDJM, LEK]
(1992) n-Dimensional nonlinear psychophysics. Erlbaum. [RAMG]
Grusser, O.-J. (1979) Cat ganglion-cell receptive fields and the role of
horizontal cells in their generation. In: The neurosciences. Fourth Study
Program, ed. F. O. Schmitt & F. G. Worden. MIT Press. [O-JG]
(1989) Quantitative visual psychophysics during the period of European
enlightenment. The studies of the astronomer and mathematician Tobias
Mayer (1723-1762) on visual acuity and colour perception. Docunwnta
Ophthalmologica 71:93-111. [O-JG]
Grusser, O.-J., Schaible, D. & Vierkant-Glathe, J. (1970) A quantitative
analysis of the spatial summation of excitation within the receptive field
centers of retinal neurons. Pfliigers Achiv der gesamten Physiologic
319;101-21. [O-JG]
Guilford, J. P. (1936) Psychometric methods. McGraw-Hill. [GB]
Gundlach, H. (1988a) Fechner scholarship. In: C. T. Fechner and psychology,
ed. J. Brozek & H. Gundlach. Passavia University Press. [HG]
(1988b) Index psychophysicus. Passavia University Press. [HG]
(1992) Entstehung und Cegenstand der Psychophysik. (Formation and object
of psychophysics.) Springer-Verlag, in press. [HG]
Haber, R. N. (1983) The impending demise of the icon: A critique of the
concept of iconic storage in visual information processing. Behavioral and
Brain Sciences 6:1-11. [rDJM]
Haller, A. von (1759) Elementa physiologiae corporis humani. Tomus quintus.
Sensus externi interni. Lausanne: Grasset. [O-JG]
(1768) Anfangsgrunde der Phisiologie des menschlichcn Korpers, vol. 5.
(Translated by J. S. Hallen.) Berlin: Voss. [O-JG]
Harnad, S., ed. (1987) Categorical perception: The groundwork of cognition.
Cambridge University Press. [aDJM]
Harper, R. & Stevens, S. S. (1948) A psychological scale of weight and a
formula for its derivation. American Journal of Psychology 61:343-
51. [MO]
Hartley, D. (1967) Thoughts about the human being, his duty and trust in
Good. In: English materialistic scientists of the eighteenth century.
Academy of the Pedagogical Sciences of USSR (in Russian). [ANL]
Harvey, O. J. & Campbell, D. T. (1963) Judgments of weight as affected by

adaptation range, adaptation duration, magnitude of unlabelled anchor,
and judgmental language. Journal of Experimental Psychology 65:12-
21. [rDJM]
Hecht, S. (1928) The relation between visual acuity and illumination. Journal
o/Ceneral Physiology 11:255-81. [O-JG]
Heidelberger, M. (1986) Zur Philosophic der Messung im 19. Jahrhundert.
In: Die historische Metrologie in den Wissenschaften, ed. H. Witthoft et
al. Scripta Mercaturae Verlag. [MH]
(1987) Fechner's indeterminism: From freedom to laws of chance. In: The
probabilistic revolution, vol. 1, Ideas in History, ed. L. KrOger, L. J.
Daston & M. Heidelberger. MIT Press. [arDJM, GG]
(1988) Fechner's Leib-Seele-Theorie. In: G. T. Fechner and psychology, ed.
J. Brozek & H. Gundlach. Passavia Universitatsverlag. (MH]
(1993a) Die innere Seite der Natur: Custav Theodor Fechners
wissenschaftlich-philosophische Weltauffassung. Vittorio Klostermann (in
press). [arDJM, MH]
(1993b) Gustav Theodor Fechner's conception of the unity of nature and
mind. In: Romantic sciences and unity of knowledge: Science 1790-1840,
ed. M. Bossi & S. Poggi. Kluwer (in press). [MH]
Heinemann, E. C. (1955) Simultaneous brightness induction as a function of
inducing- and test-field luminances. Journal of Experimental Psychology
50:89-96. [aDJM]
Heller, O. (1990) Scaling and orientation. In: Fechner Day 90 - Proceedings
of the Sixth Annual Meeting of the International Society for
Psychophysics, ed. F. Miiller & F. R. G. Wuraburg. [GB]
Helmholtz, H. von (1856-66/1962) Treatise on physiological optics.
(Translated by J. P. C. Southall.) Dover. [aDJM]
(1866) Handbuch der physiologischen Optik. Leipzig: Voss. [O-JG]
(1887/1921) Zfihlen und Messen, erkenntnistheoretisch betrachtet. In:
Schriften zur Erkenntnistheorie, ed. P. Hertz & M. Schlick. Julius
Springer. [MH]
Henmon, V. A. C. (1906) The time of perception as a measure of differences
in sensation. Archives of Philosophy, Psychology and Scientific Methods
8:1-75. [WMP]
Herbart, J. F. (1812) Psychologische Untersuchung ueber die Staerke einer
gegebenen Vorstellung, als Function ihrer Dauer betrachtet. In: Johann
Friderich Herbart's Schriften zur Psychologic, Drifter Theil, ed. G.
Hartenstein (1989). Leipzig: Voss. [rDJM, RAMC]
(1816/1891) Lehrbuch der Psychologie. (Translated by M. K. Smith as: A
textbook in psychology.) D. Appleton. [rDJM]
(1824/1892) Psychologie als Wissenschaft neu gegriindet auf Erfahrung,
Metaphysik und Mathematik. In: J. F. Herbart's samtliche Werke Part 1
(vol. 5) & Part 2 (vol. 6) ed. K. Kehrbach. Langensalza: H. Beyer &
S6hne. [rDJM]
Hering, E. (1874) Zur Lehre vom Lichtsinn V. Grundziige einer Theorie des
Lichtsinnes. Sitzungsberichte der Kaiserlichen Akademie der
Wissenschaften in Wien. Mathematisch-Noturwissenschaftliche Classe,
Abth. Ill 69:85-104. [O-JG]
(1875) Zur Lehre von der Beziehung zwischen Leib und Seele. I.
Mittheilung. Ober Fechner's psychophysisches Gesetz. Sitzungsberichte
der Kaiserlichen Akademie der Wissenschaften, Wien. Mathematisch-
Natunvisscnschaftliche Klasse. Abth. Ill 72:310-48. [aDJM, MH]
(1920) Crundziige der Lehre com Lichtsinn. Springer. [O-JG]
(1931) Wissenschaftliche Abhandlungen, 2 vols. Thiem. [O-JG]
Hess, C. & Pretori, H. (1894) Messende Untersuchungen uber die
Gesetzmassingkeit des simultanen Helligkeits-Contrastes. Archiv fur
Ophthalmologie 40:1-27. (Translated [1969] by H. R. Flock & J. H.
Tenney). Technical Report FLP.-l, York University. [aDJM]
Hick, VV. E. (1952) On the rate of gain of information. Quarterly Journal of
Experimental Psychology 4:11-26. [rDJM]
Hoffman, W. C. & Dodwell, P. C. (1985) Geometric psychology generates the
visual Cestalt. Canadian Journal of Psychology 39:491-528. [rDJM]
Hooke, R. (1705) The posthumous works. London: Smith & Walford. [O-JG]
Hornstein, G. A. (1988) Quantifying psychological phenomena: Debates,
dilemmas, and implications. In: The rise of experimentation in American
psychology, ed. J. G. Morawski. Yale University Press. [GAH]
Houtsma, A. J. M., Durlach, N. I. & Braida, L. D. (1980) Intensity
perception XI. Experimental results on the relation of intensity resolution
to loudness matching. Journal of the Acoustical Society of America
68:807-13. [LEK]
Hueck, A. (1840) Ober die Grenzen des Sehvermogens. Miillers Archie fur
Anatomic, Physiologic und Wissenschaftliche Medizin 1840:82-97.
[O-JG]
Hurvich, L. & Jameson, D. (1966) The perception of brightness and darkness.
Allyn & Bacon. [aDJM]
lllich, 1. (1985) H2O and the toaters of forgetfulness. Heyday Books.
[DAS]
References/Murray: History of psychophysics
Indow, T. (1989) Psychophysics: On the possibility of another approach.
Behavioral and Brain Sciences 12:276-77. [aDJM]
Jackson, S. W. (1986) Melancholia and depression from Hippocratic times to
modern times. Yale University Press. [rDJM]
James, W. (1890) Principles of psychology. Holt, Rinehart & Winston.
(Reprint 1950, Dover) [rDJM, DAS, MW]
(1911) The will to believe and other essays in popular philosophy. Longman,
Green. [LEM]
Jameson, D. & Hurvich, L. M. (1961) Complexities of perceived brightness.
Science 133:174-79. [aDJM]
Jarvilehto, T. (1981) A new combination of methods for the study of neural
mechanisms of mental activity. In: Soviet-Finnish symposium on
psychophysiology, Moscow. Finnish-Soviet Committee on Scientific
Technical Cooperation. [ANL]
Jastrow, J. (1888) A critique of psycho-physic methods. American Journal of
Psychology 1:271-309. [aDJM]
Jesteadt, W., Wier, C. C. & Green, D. M. (1977) Intensity discrimination as
a function of frequency and sensation level. Journal of the Acoustical
Society of America 61:169-77. [aDJM]
Jones, L. A. (1986) Perception of force and weight: Theory and research.
Psychological Bulletin 100:29-42. [HER]
Kaas, J. (1989) The segregation of function in the nervous system: Why do
sensory systems have so many subdivisions? In: Contributions to sensory
physiology, vol. 7, ed. W. D. Neff. Academic Press. [SH]
Kaestner, A. G. (1762) Elogium Tobiae Mayeri. Gottingen: Schulz &
Rosenbusch. [O-JG]
Katz, D. (1925/1989) The world of touch. (Translated and ed. by L. E.
Krueger.) Erlbaum. [aDJM]
Kim, D. O., Chang, S. O. & Sirianni, J. G. (1980) A population study of
auditory nerve fibers in unanesthetized decerebrate cats: Response to
pure tones. Journal of the Acoustical Society of America 87(4):1648-
55. [SH]
Kimchi, R. (1992) Primacy of holistic processing and global/local paradigm: A
critical review. Psychological Bulletin 112:24-38. [rDJM]
Knibestol, M. & Vallbo, A. B. (1980) Intensity of sensation related to activity
of slowly adapting mechanoreceptive units in the human hand. Journal of
Physiology 300:251-67. [aDJM]
Kohler, A. (1886) Ueber die hauptsachlichsten Versuche einer
mathematischen Formulirung des psychophysischen Gesetzes von Weber.
Philosophische Studien 3:572-642. [aDJM]
Kohler, W. (1969) The task of Cestalt psychology. Princeton University
Press. [rDJM]
Konig, A. (1897) Die Abhangigkeit der Sehscharfe von der
Beleuchtungsintensitat. Berichte der Koniglich Preussischen Akademie
der Wissenschaften 1897:559-75. [O-JG]
Krantz, D. H. (1969) Threshold theories of signal detection. Psychological
Review 76:308-24. [aDJM, DL]
(1972) Visual scaling. In: Visual psychophysics, vol. 7/4, Handbook of
sensory physiology, ed. D. Jamson & L. E. Hurwich. Springer. [O-JG]
Krantz, D. H., Luce, R. D., Suppes, P. &Tversky, A. (1971) Foundations of
measurement, vol. 1. Academic Press. [RDL, WMP]
Kries, J. von (1882) Ober die Messung intensiver Grossen und fiber das
sogenannte psychophysische Gesetz. Vierteljahrsschrift fur
wissenschaftliche Philosophic 6:257-94. [arDJM, MH, JVB]
(1886) Prinzipien der Wahrscheinlichkeitsrechnung. Mohr. [MH]
Krueger, L. E. (1989a) Reconciling Fechner and Stevens: Toward a unified
psychophysical law. Behavioral and Brain Sciences 12:251-320. [aDJM,
JVB, LEK, DJW]
(1989r) Psychophysical law: Keep it simple. Behavioral and Brain Sciences
12:299-320. [aDJM]
(1991) Psychophysical law: Taming the cognitive and chaotic aspects.
Behavioral and Brain Sciences 14:193-204. [aDJM, LEK)
Kruskal, J. B. (1964) Nonmetric multidimensional scaling: A numerical
method. Psychometrika 29:115-29. [WMP]
Kuhl, P. K., Williams, K. S., Lacerda, F., Stevens, K. N. & Lindblom, B.
(1992) Linguistic experience alters phonetic perception in infants by 6
months of age. Science 255(5044):606-08. [SH]
Kiilpe, O. (1893) Grvndriss der Psychologie. Herzil. [RAMC]
LaMettrie, J. O. (1751) L'homme machine. English translation in: Man a
machine [1912], ed. C. C. Bussey. Open Court. [O-JG]
Laming, D. (1984) The relativity of "absolute' judgments. British Journal of
Mathematical and Statistical Psychology 37:152-83. [aDJM]
(1985) Some principles of sensory analysis. Psychological Review 92:462
85. [aDJM]
(1986) Sensory analysis. Academic Press. [aDJM, DL]
(1988) A reexamination of Sensory analysis. Behavioral and Brain Sciences
11:316-39. [rDJM]
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 181
 References/Murray: History of psychophysics
(1989) Experimental evidence for Fechner's and Stevens's Laws. Behavioral
and Brain Sciences 12:277-81. [aDJM, RTl
(1991) Reconciling Fechner and Stevens? Behavioral and Brain Sciences
14:188-91. [aDJM, MO]
Leary, D. E. (1980) The historical foundation of Herbart's mathematization of
psychology. Journal of the Behavioral Sciences 16:150-63. [rDJM]
Lebedev, A. N. (1990) Cyclic neural codes of human memory and some
quantitative regularities in experimental psychology. In: Psychophysical
explorations of mental structures, ed. H.-G. Geissler. Hogrefe &
Huber. [ANL]
Lebedev, A. N. & Myshkin, I. Yu. (1989) Contemporary problems of inner
psychophysics. In: Fechner Day 90. Proceedings of the Sixth Annual
Meeting of the International Society of Psychophysics, ed. F. Muller.
Universitat Wuerzburg. [ANL]
Legge, G. E. (1978) Sustained and transient mechanisms in human vision:
Temporal and spatial properties. Vision Research 18:69-81. [DL]
Legge, G. E. & Foley, J. M. (1980) Contrast masking in human vision.
Journal of the Optical Society of America 70:1458-71. [aDJM]
Lehmann, A. (1886) Ueber die Anwendung der Methode der mittleren
Abstufungen auf den Lichtsinn. Philosophische Studien 3:497
533. [aDJM]
Lewin, K. (1936) Principles of topological psychology. McGraw-Hill. [LEM]
Libet. B.. Alberts, W. W., Wright. E. W. Jr. & Feinstein, B. (1967)
Responses of human somatosensory cortex to stimuli below threshold for
consciousness. Science 158:1597-1600. [aDJM]
Lichtenberg, G. C. (1775) Tobiae Mayeri opera inedita I. Gottingen:
Vandenhoeck. [O-JG]
Lim, J. S., Rabinowitz, W. M., Braida, L. D. & Durlach, N. I. (1977)
Intensity perception VIII. Loudness comparisons between different types
of stimuli. Journal of the Acoustical Society of America 62:1256-67.
Erratum: (1978) 63:1233. [LEK]
Link, S. W. (1992) The wave theory of difference and similarity.
Erlbaum. [rDJM, SL]
Lipetz, L. E. (1971) The relation of physiological and psychological aspects of
sensory intensity. In: Handbook of sensory physiology, vol. 1, ed. W. R.
Loewenstein. Springer-Verlag. [arDJM, LEK]
Lockhead, G. R. (1992) Psychophysical scaling: Judgments of attributes or
objects? Behavioral and Brain Sciences 15:543-601. [GRL]
Lotze, H. (1887) Metaphysic, vol. 2 (Translated by B. Bosanquet). Clarendon
Press. [DAS]
Lowrie, W., trans, and ed. (1946) Religion of a Scientist; Selections from
Gustao Th. Fechner. Pantheon Books. [rDJM]
Luce, R. D. (1959) On the possible psychophysical laws. Psychological Review
66:81-95. [DJW]
(1963) A threshold theory for simple detection experiments. Psychological
Review 70:61-79. [aDJM, DL]
(1972) What sort of measurement is psychophysical measurement? American
Psychologist 27:96-106. [aDJM]
(1977) A note on sums of power functions. Journal of Mathematical
Psychology 16:91-93. [RDL]
(1986) Response times: Their role in inferring elementary mental
organization. Oxford University Press. [rDJM, RDL]
Luce, R. D. & Edwards, W. (1958) The derivation of subjective scales from
just noticeable differences. Psychological Review 65:222-37. [aDJM]
Luce, R. D., Krantz, D. H., Suppes, P. & Tversky, A. (1990) Foundations of
measurement, vol. 3. Academic Press. [RDL]
Luce, R. D. & Narens, L. (1987) Measurement scales on the continuum.
Science 235:1527-32. [aDJM, DJW]
Mach, E. (1866) Uber den psychologischen Effect raumlich vertheilter
Lichtreize. Sitzungsberichte der Kaiserlichen Akademie der
Wissenschaften in Wien. Mathematisch-Naturwissenschafliche Klasse,
Abth II. 54:131-44. [rDJM]
(1868) Uber die physiologische Wirkung raumlich vertheilter Lichtreize.
Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften, in Wien.
Mathematisch-Naturwissenschoftliche Klasse Abth U 57:11-19. [O-JG,
MH]
(1883/1960) The science of mechanics: A critical b historical account of its
development, 6th ed. (after the 9th German edition). Open Court. [JVB]
(1886/1914) The analysis of sensations and the relation of the physical to the
psychical. (Translated by C. M. Williams & S. Waterlow.) Open
Court. [aDJM]
(1886/1919) Die Analyse der Empflndungen und des Verhaltnis des
Physischen zum Psychischen, 8th ed. Gustav Fischer. [JVB]
(1886/1922) Die Analyse der Empfindungen und das Verhaltnis des
Physischen zum Psychischen, 9th ed. Gustav Fischer. [MH]
(1891) Some questions of psycho-physics. The Monist 1:393-400. [rDJM]
(1896/1919) Die Principled der Warmelehre. Historisch-kritisch entwickelt,
3rd ed. Johann Ambrosius Barth. [MH]
182 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
Macmillan, N. A. (1987) Beyond the categorical/continuous distinction: A
psychophysical approach to processing modes. In: Categorical Perception,
ed. S. Hamad. Cambridge University Press. [aDJM]
Macmillan, N. A. & Creelman, C. D. (1991) Detection theory: A user's guide.
Cambridge University Press. [arDJM, PW]
Marcel, A. J. (1983) Conscious and unconscious perception: Experiments on
visual masking and word recognition. Cognitive Psychology 15:197
237. [aDJM]
Markowitz, J. & Swets, J. A. (1967) Factors affecting the slope of empirical
ROC curves: Comparison of binary and rating responses. Perception and
Psychophysics 2:91-150. [aDJM]
Marks, L. E. (1974) On scales of sensation: Prolegomena to any future
psychophysics that will be able to come forth as science. Perception and
Psychophysics 16:358-76. [aDJM]
(1979) A theory of loudness and loudness judgments. Psychological Review
86:256-85. [aDJM]
Marks, L. E., Borg, G. & Ljunggren, G. (1983) Individual differences in
perceived exertion assessed by two new methods. Perception and
Psychophysics 34:280-88. [GB]
Marks, L. E., Szczesiul, R. & Ohlott, P. (1986) On the cross-modal
perception of intensity. Journal of Experimental Psychology: Human
Perception and Performance 12:517-34. [aDJM]
Marshall, M. E. (1982) Physics, metaphysics and Fechner's psychophysics. In:
The problematic science: Psychology in nineteenth century thought, ed.
W. R. Woodward & M. G. Ash. Praeger. [rDJM]
(1990) The theme of quantification and the hidden Weber in the early work
of Gustav Theodor Fechner. Canadian Psychology 31:45-53. [DAS]
Massaro, D. W. (1987) Categorical partition: A fuzzy-logical model of
categorization behavior. In: Categorical perception, ed. S. Hamad.
Cambridge University Press. [aDJM]
Matthews, B. H. C. (1931) The response of a muscle spindle during active
contraction of a muscle. Journal of Physiology 72:153-74. [rDJM]
Mayer, T. (1741) Neue und allgemeine Art, alle Aufgaben aus der Geometrie
vermittelst der geometrischen Linien leichter aufzulosen; insbesondere
wie alle regulare und irregulare Vielecke, davon ein Verhaltnis ihrer
Seiten gegeben in den Circul geometrisch sollen eingeschrieben werden,
samt einer kurzen hierzu notigen Buchstaben-Rechenkunst und
Geometrie. Als Erstlinge an das Licht gestellet von Tobias Mayern.
Esslingen: Mantlern. [O-JG]
(1745) Mathematischer Atlas, in welchem auf 60 Tabellen alle der
Mathematik vorgestellt, und nicht allein iiberhaupt zu bequemer
Wiederholung, sondern auch den Anfangern besonders zur
Aufmunterung durch deutliche Beschreibung und Figuren entworfen
werden. Augsburg: Pfeffel. [O-JG]
(1754a) Observationes astronomicae A. 1753 Gottingae habitae. Commentarii
Societatis Regiae Scientiarium Cottingensis 3:441-51. [O-JG]
(1754b) Experimente uber die Scharfe des Gesichtsinnes. Cottingische
Anzeigen von gelehrten Sachen (20):401-2. [rDJM, O-JG]
(1755) Experimenta circa visus aciem. Commentarii Societatis Regiae
Scientiarium Cottingensis 4:97-112. [rDJM, O-JG]
(undated, ca. 1757) Von den Farben. Unpublished manuscript, Gottingen
University Library, Mayer 15, 13. [O-JG]
(1758) De affinitate colorum (Uber die Verwandtschaft der Farben).
Cottingische Anzeigen von gelehrten Sachen, 147. Stuck (9
Dezember): 1385-89. [O-JG]
(1759) Die verschiedenen Grade des Schattens zu bestimmen, wenn man
eine einzige Farbe zugleich zum ganzen und halben Schatten nehmen
will. Unpublished manuscript January 22, 1759, Gottingen University
Library, Mayer 13. [O-JG]
(1775) De affinitate colorum. In: Tobiae Mayeri opera inedita /, ed. G. C.
Lichtenberg. Gottingen: Vandenhoeck. (English translation in Forbes
1971.) [O-JG]
McBride, R. L. & Anderson, N. H. (1991) Integration psychophysics in the
chemical senses. In: Contributions to information integration theory, vol.
1, Cognition, ed. N. H. Anderson. Erlbaum. [NHA]
McElree, B. & Dosher, B. A. (1989) Serial position and set-size in short-term
memory: The time course of recognition. Journal of Experimental
Psychology: General 118:346-73. [rDJM]
McKenna, F. P. (1985) Another look at the "new psychophysics." British
Journal of Psychology 76:97-109. [HER]
McNicol, D. (1972) A primer of signal detection theory. George Allen &
Unwin. [aDJM)
McNicol, D. & Stewart, G. W. (1980) Reaction time and the study of
memory. In: Reaction times, ed. A. T. Welford. Academic
Press. [rDJM]
McReynolds, P. & Ludwig, K. (1977) Psychometrics in the seventeenth
century: The personology of Christian Thomasius. Paper presented at the
9th Annual Meeting of Cheiron, the International Society for the History

of the Behavioral and Social Sciences, University of Colorado, Boulder,
June 9-11. [rDJM]
Melara, R. D. & Marks, L. E. (1990) Interaction among auditory dimensions:
Timbre, pitch, and lotidncss. Perception and Psychophysics 48(2):169
78. [SH]
Melzack, R. (1992) Phantom limbs. Scientific American 266(4):120-
26. [WLB]
Merzenich, M. M., Jenkins, W. M. & Middlebrooks, J. C. (1984)
Observations and hypotheses on special organizational features of the
central auditory nervous system. In: Dynamic aspects of neocortical
function, ed. G. M. Edelman, W. E. Gall & W. M. Cowan.
Wiley. [SH]
Mctraux, A. (1983) An essay on the early beginnings of psychometrics. In:
Advances in historiography of psychology, ed. G. Eckart & L. Sprung.
VEB Deutscher Verlag der Wissenschaften. [rDJM]
Michcll, J. (1986) Measurement scales and statistics: A clash of paradigms.
Psychological Bulletin 100:398-407. [MH]
Michels, W. C. 6t Helson, H. (1949) A reformulation of the Fechner Law in
terms of adaptation-level applied to rating-scale data. American Journal of
Psychology 62:355-68. [aDJM]
Mises, R. von (1928/1957) Probability, statistics, and truth. Dover. [GG]
Molcn, M. W. van der & Keuss, P. J. C. (1979) The relationship between
reaction time and auditory intensity in discrete auditory tasks. Quarterly
Journal of Experimental Psychology 31:95-102. [rDJM]
Montgomery, H. (1975) Direct estimation: Effect of methodological factors on
scale type. Scandinavian Journal of Psychology 16:19-29. [aDJM, MO]
Moravec, H. (1988). Mind children: The future of robot and human
intelligence. Harvard University Press. [MH]
Morgan, C. L. (1900) On the relation of stimulus to sensation in visual
impressions. Psychological Review 7:217-33. [aDJM]
Mountcastle, V. G., Poggio, G. F. & Werner, G. (1963) The relation of
thalamic cell response to peripheral stimuli varied over an intensive
continuum. Journal of Neurophysiology 26:80734. [GB]
Munstcrberg, H. (1894) Studies from the Harvard psychological laboratory: A
psychometric investigation of the psycho-physic law. Psychological Review
1:45-51. [WMP]
Miiller, G. E. (1878) Zur Grundlcgung der Psychophysik. Berlin: T.
Grieben. [aDJM]
Murray, D. J. (1967) The role of speech responses in short-term memory.
Canadian Journal of Psychology 21:263-76. [aDJM]
(1975) Craphemically cued retrieval of words from long-term memory.
Journal of Experimental Psychology: Human Learning and Memory
104:65-70. [rDJM]
(1985) Review of Tulving, E.: Elements of episodic memory. Canadian
Psychology 26:235-38. [aDJM]
(1986) Level of articulation and short-term recognition following brief probe
delays. Bulletin of the Psychonomic Society 24:103-6. [rDJM]
(1987) A perspective for viewing the integration of probability theory into
psychology. In: The probabilistic revolution, vol. 2, Ideas in the sciences,
ed. L. Kriiger, C. Cigerenzer & M. S. Morgan. MIT Press. [arDJM,
GG]
(1988) A history of Western psychology, 2nd ed. Prentice-Hall. [rDJM]
(1989) Nineteenth-century attempts to decide between psychophysical laws.
Behavioral and Brain Sciences 12:284-85. [aDJM]
(1990) Fechner's later psychophysics. Canadian Psychology 31:54-
60. [aDJM]
Murray, D. J. & Ross, H. E. (1988) E. H. Weber and Fechner's
psychophysics. In: G. T. Fechner and psychology, ed. J. Brozek & M.
Gundlach. Passavia UniversitStsverlag. [aDJM]
Myers, A. K. (1982) Psychophysical scaling and scales of physical stimulus
measurement. Psychological Bulletin 82:303-14. [LEK, DJW]
Naka, K.-I. & Rushton, W. A. H. (1966) S-potentials from luminosity units in
the retina offish (Cyprinidae). Journal of Physiology (London) 185:587-
99. IO-JG]
Narens, L. (1992) A theory of magnitude estimation. Submitted for
publication. [RDL]
Neisser, U. (1982) Memory observed: Remembering in natural contexts.
W. H. Freeman. [rDJM]
Newell, A. (1990) Unified theories of cognition. Harvard University
Press. [rDJM]
Newhall, S. M., Nickerson, D. & Judd, D. B. (1943) Final report of the
O.S.A. subcommittee on the Munsell colors. Journal of the Optical
Society of America 33:385-418. [aDJM]
Newman, E. B. (1974) On the origin of scales of measurement. In: Sensation
and measurement, ed. H. R. Moskowitz et al. Reidel. [MH]
Neyman, J. (1976) The emergence of mathematical statistics. In: On the
history of statistics and probability, ed. D. B. Owen. Marcel
Dekker. [SL]
References /Murray: History of psychophysics
Neyman, J. & Pearson, E. S. (1928) On the use and interpretation of certain
test criteria for purposes of statistical inference. Part I. Biometrika
20A:175-240, 263-294. [aDJM, DL]
(1933) On the problem of the most efficient tests of statistical hypotheses.
Philosophical Transactions of the Royal Society of London 231A:289-
337. [DL]
Norman, D. A. (1964) Sensory thresholds, response biases, and the neural
quantum theory. Journal of Mathematical Psychology 1:88-120. [DL]
Norwich, K. H. (1987) On the theory of Weber fractions. Perception and
Psychophysics 42:286-98. [aDJM, DJW]
(1991) Toward the unification of the laws of thought: Some food for thought.
In: Sensory science theory and applications in foods, ed. H. T. Lawless 6t
B. P. Klein. Marcel Dekker. [rDJM]
Nunnally, J. C. (1970) Introduction to psychological measurement. McGraw-
Hill. [DJW]
Parducci, A. (1963) Range-frequency compromise in judgment. Psychological
Monographs 77 (2, whole no. 565). [rDJM]
(1982) Scale values and phenomenal experience: There is no psychophysical
law! In: Psychophysical judgment and the process of perception, ed.
H. G. Geissler, P. Petzold, H. F. J. M. Buffart & Y. M. Zabrodin. VEB
Deutscher Verlag der Wissenschaften. [GG]
(1984) Category ratings and the relational character of judgement. In:
Modern trends in perception, eds. H. G. Geissler & V. Sarris. VEB
Deutscher Verlag der Wissenschaften. [GB]
Parker, S. & Schneider, B. (1980) Loudness and loudness discrimination.
Perception and Psychophysics 28:398-406. [aDJM]
Parker, S., Schneider, B. & Kanow, G. (1975) Ratio scale measurement of the
perceived lengths of lines. Journal of Experimental Psychology: Human
Perception and Performance 104:195-204. [aDJM]
Parks, T. R. (1966) Signal detectability theory of recognition-memory
performance. Psychological Review 73:44-58. [aDJM]
Pascal, B. (1670/1958) Penstes. Dutton. [LEM]
Pastore, R. D. (1987) Categorical perception: Some psychophysical models.
In: Categorical perception, ed. S. Harnad. Cambridge University
Press. [aDJM, SH]
Patton, H. D. (1960) Reflex regulation of movement and posture. In: Medical
physiology and biophysics, ed. T. C. Ruch & J. F. Fulton. W. B.
Saunders. [rDJM, WLB]
Peterson, W. W., Birdsall, T. G. & Fox, W. C. (1954) The theory of signal
detectability. Transactions of the IRE Professional Croup in Information
Theory PCIT 2-4:171-212. [DL]
Petrusic, W. M. & Jamieson, D. J. (1979) Resolution time and the coding of
arithmetic relations on supraliminally different visual extents. Journal of
Mathematical Psychology 19:89-107. [WMP]
Pieron, H. (1914) Recherches sur les lois de variation des temps de latence
sensorielle en fonction des intensites excitatrices. VAnnie Psychologique
20:17-96. [ANL]
Plateau, J. A. F. (1872a) Sur la mesure des sensations physiques, et sur la loi
qui lie I'intensit6 de ces sensations a 1'intensit^ de la cause excitante.
Bulletins de I'Acad&nie Royale de Belgique (2nd series) 33:376-
88. [aDJM, O-JG]
(1872b) Un mot au suject du mfimoire presente par M. Delboeuf. Bulletins
de I'Acadtmie Royale de Belgique (2nd series) 34:141-42. [aDJM]
(1873) Uber die Messung physischer Empfindungen und das Gesetz,
welches die Starke dieser Empfindungen mit der Starke der erregende
Ursache verknupft. Poggendorfs Annalen der Physik und Chemie
(5)30:465-76. (O-JG]
Plato (1899a) Protagoras. In: The dialogues of Plato, vol. 1. (Translated and
ed., B. Jowett.) Scribner's. [LEM]
(1899b) The republic. In: The dialogues of Plato, vol. 2 B. Jowett (trans.
and ed.). Scribner's. [LEM]
Popper, R., Parker, S. & Galanter, E. (1986) Dual loudness scales in
individual subjects. Journal of Experimental Psychology: Human
Perception and Performance 12:61-69. [aDJM]
Poulton, E. C. (1968) The new psychophysics: Six models for magnitude
estimation. Psychological Bulletin 69:1-19. [aDJM, HER]
(1989) Bias in quantifying judgments. Erlbaum. [rDJM, HER]
Priest, I. G., Gibson, K. S. & McNicholas, H. J. (1920) An examination of the
Munsell color system. I. Spectral and total reflection and the, Munsell
scale of value. (U.S. National Bureau of Standards Technical Report 167).
U.S. National Bureau of Standards. [aDJM]
Putnam, H. (1967/1985) The nature of mental states. In: Mind, language and
reality: Philosophical papers, vol. 2, ed. H. Putnam. Cambridge
University Press. [MH]
Pylyshyn, Z. W. (1984) Computation and cognition: Toward a foundation for
cognitive science. MIT Press (Bradford Books). (aDJM]
Quick, R. F. (1974) A vector-magnitude model of contrast detection.
Kybernetik 16:65-67. [DL]
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 183
References/Murray: History of psychophysics
Quine, W. V. (1960) Word and object. MIT Press. [GB]
(1987) Quiddities. Harvard University Press. [GB]
(1990) Pursuit of truth. Harvard University Press. [GB]
Ramul, K. (1960) The problem of measurement in the psychology of the
eighteenth century. American Psychologist 15:256-65. [rDJM]
(1963) Some early measurements and ratings in psychology. American
Psychologist 18:653-59. [rDJM]
Ratliff, F. & Hartline, H. K. (1959) The response of Limulus optic nerve
fibers to patterns of illumination on the receptor mosaic. Journal of
General Physiology 42:1241-55. [rDJM, WLB]
Rice, S. O. (1944/1945) Mathematical analysis of random noise. Bell System
Technical Journal 23:282-332 & 24:46-156. [DL]
Richter, M. (1953) Das System der DIN-Farbenkarte. Die Farbe 1:85-
98. [aDJM]
Riesz, R. R. (1933) The relationship r ;tween loudness and the minimum
perceptible increment of intens.cy. Journal of the Acoustical Society of
America 4:211-16. [LEK]
Robson, J. G. & Graham, N. (1981) Probability summation and regional
variation in contrast sensitivity across the visual field. Vision Research
21:409-18. [DL]
Rosen, S. & Howell, P. (1987) Auditory, articulatory, and learning
explanations of categorical preception in speech. In: Categorical
perception, ed. S. Hamad. Cambridge University Press. [SH]
Roskam, E. E. (1991) Psychophysical scaling. In: Fechner Day 91. Proceedings
of the Seventh Annual Meeting of the International Society for
Psychophysics, ed. G. R. Lockhead. Duke University, Durham,
NC. [aDJM]
Ross, H. E. (1969) When is a weight not illusory? Quarterly Journal of
Experimental Psychology 21:346-55. [HER]
Rozsypol, A. J., Stevenson, D. C. & Hogan, J. T. (1985) Dispersion in models
of categorical perception. Journal of Mathematical Psychology 29:271-
88. [aDJM]
Rule, S. J. & Curtis, D. W. (1977) The influence of the interaction of weight
and volume on subjective heaviness. Perception and Psychophysics
22:159-64. [HER]
Rule, S. J., Curtis, D. W. & Mullin, L. C. (1981) Subjective rates and
differences in perceived heaviness. Journal of Experimental Psychology:
Human Perception and Performance 7:459-66. [aDJM]
Rule, S. J., Laye, R. C. & Curtis, D. W. (1974) Magnitude judgments and
difference judgments of tightness and darkness: A two-stage analysis.
Journal of Experimental Psychology 103:1108-14. [aDJM]
Rumelhart, D. E., McClelland, J. L. & the PDP Research Group (1986)
Parallel distributed processing: Explorations in the microstructure of
cognition, vol. 1, Foundations. MIT Press. (aDJM]
Sakitt, B. (1974) Canonical ratings. Perception and Psychophysics 16:478-
88._ [aDJM]
Sarris, V. (1967) Adaptation level theory: Two critical experiments on Helson's
weighted average model. American Journal of Psychology 80:331-
55. [ANL]
Savage, C. W. (1970) The measurement of sensation. University of California
Press. [aDJM]
Scheerer, E. (1987) The unknown Fechner. Psychological Research 49:197-
202. [aDJM]
(1989) Conjuring Fechner's spirit. Behavioral and Brain Sciences 12:288-
90. [DAS]
(1992) Fechner's inner psychophysics: Its historical fate and present status.
In: Cognition, information processing and psychophysics: Basic issues,
ed. H.-G. Geissler, S. W. Link & J. T. Townsend. Erlbaum. [aDJM,
SL]
Scheerer, G. (1987) Tobias Mayer - experiments on visual acuity. Spatial
Vision 2:81-97. [O-JG]
Schmidt, S. R. (1991) Can we have a distinctive theory of memory? Memory
and Cognition 19:523-42. [aDJM]
Schneider, B. (1980a) Individual loudness functions determined from direct
comparisons of loudness intervals. Perception and Psychophysics 28:493-
503. [aDJM]
(1980b) A technique for the nonmetric analysis of paired comparisons of
psychological intervals. Psychometrika 45:357-72. [aDJM]
(1989) Is there really only one representation for stimulus intensity?
Behavioral and Brain Sciences 12:290. [DJW]
Schneider, B. & Bissett, R. (1988) "Ratio" and "difference" judgments for
length, area, and volume: Are there two classes of sensory continua?
Journal of Experimental Psychology: Human Perception and Performance
14:503-12. [aDJM]
Schneider, B., Parker, S. & Upenieks, E. G. (1982) The perceptual basis of
judgments of pitch differences and pitch ratios. Canadian Journal of
Psychology 36:4-23. [aDJM]
184 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1
Shannon, C. E. (1949) Communication in the presence of noise. Proceedings
of the Institute of Radio Engineers 37:10-21. [DL]
Shepard, R. N. (1962) Analysis of proximities: Multidimensional scaling with
an unknown distance function. Psychometrika 27:125-40. [WMP]
(1981) Psychological relations and psychological scales: On the status of
"direct" psychophysical measurement. Journal of Mathematical
Psychology 24:21-57. [arDJM]
Shipley, T., ed. (1961) Classics in psychology. Philosophical Library. [rDJM]
Smith, P. T. (1991) Is this effort worth the pain? British Journal of
Mathematical and Statistical Psychology 44:413-17. [RAMC]
Solomons, L. M. (1900) A new explanation of Weber's Law. Psychological
Review 7:234-40. [aDJM]
Sprung, L. & Sprung, H. (1987) Gustav Theodor Fechner in der Ceschichte
der Psychologie - Leben, Werk und Wirken in der
Wissenschafisentwicklung des 19 Jahrhunderts. Karl-Marx
UniversitSt. [RAMG]
Star, S. L. (1985) Scientific work and uncertainty. Social Studies of Science
15:391-427. [GAH]
(1986) Triangulating clinical and basic research: British localizationists,
1870-1906. History of Science 24:29-48. [GAH]
(1989) Regions of the mind: Brain research and the quest for scientific
certainty. Stanford University Press. [GAH]
Stemberg, S. (1966) High speed scanning in human memory. Science
153:652-54. [rDJM]
(1969) The discovery of processing stages: Extensions of Donders" method.
In: Attention and performance, vol. 2, ed. W. G. Koster. North-
Holland. [rDJM]
Stevens, J. C. & Rubin, L. L. (1970) Psychophysical scales of apparent
heaviness and the size-weight illusion. Perception and Psychophysics
8:225-30. [HER]
Stevens, S. S. (1936) A scale for the measurement of a psychological
magnitude: Loudness. Psychological Review 43:405-16. [aDJM]
(1946) On the theory of scales of measurement. Science 103:677-
80. [aDJM, MH]
(1951) Handbook of experimental psychology. Wiley. [aDJM]
(1957) On the psychophysical law. Psychological Review 64:153-
81. [arDJM, GB, MT, MW]
(1961) To honor Fechner and repeal his law. Science 133:80-86. [aDJM,
MT]
(1970) Neural events and the psychophysical law. Science 170:1043-
50. [arDJM]
(1971) Issues in psychophysical measurement. Psychophysical Review
78:426-50. [GB]
(1975) Psychophysics: Introduction to its perceptual, neural, and social
prospects, ed. G. Stevens. Wiley. [rDJM, MH, MO]
Stevens, S. S. & Galanter, E. H. (1957) Ratio scales and category scales for a
dozen perceptual continua. Journal of Experimental Psychology 54:377-
411. [rDJM, HER]
Stevens, S. S., Morgan, C. T. & Volkmann, J. (1941) Theory of the neural
quantum in the discrimination of loudness and pitch. American Journal of
Psychology 54:315-35. [aDJM, LEK]
Stevens, S. S. & Stevens, J. C. (1960) Brightness function: Parametric effects
of adaptation and contrast. Journal of the Optical Society of America
50:1139. [aDJM]
Stigler, S. M. (1986) The history of statistics. The measurement of uncertainty
before 1900. Belknap/Harvard University Press. [GG]
Stout, D. (1989) A question of statistical inference: E. G. Boring, T. L. Kelley
and the probable error. American Journal of Psychology 102:549-
62. [aDJM]
(in press) E. G. Boring, C. Murchison, and Clark University: Research
practice and practical politics. Recent trends in theoretical psychology,
vol. 3. Springer-Verlag. [DAS]
Stout, G. F. (1888) The Herbartian psychology. Mind 13:321-88; 473-
98. [rDJM]
Suppes, P. & Zinnes, J. L. (1963) Basic measurement theory. In: Handbook
of mathematical psychology, vol. 1, ed. R. D. Luce et al. Wiley.
[MH]
Swartz, P. (1953) A new method for scaling pain. Journal of Experimental
Psychology 45:288-93. [aDJM]
Swets, J. A. (1986a) Indices of discrimination or diagnostic accuracy: Their
ROCs and implied models. Psychological Bulletin 99:100-117. [aDJM]
(1986b) Form of empirical ROCs in discrimination and diagnostic tasks:
Implications for theory and measurement of performance. Psychological
Bulletin 99:181-98. [aDJM]
Swets, J. A., Tanner, W. P. & Birdsall, T. G. (1961) Decision processes in
perception. Psychological Review 68:301-40. [arDJM, DL]
Takane, Y., Young, F. W. & de Leeuw, J. (1977) Nonmetric individual

differences multidimensional scaling: An alternating least squares method
with optimal scaling features. Psychometrika 42:7-67. [WMP]
Takasaki, H. (1966) Lightness changes of grays induced by change in
reflectance of gray background. Journal of the Optical Society of America
56:504-9. [rDJM]
Tanner, W. P., Jr. (1956) Theory of recognition. Journal of the Acoustical
Society of America 28:882-88. [aDJM]
Tanner, W. P. & Swets, J. A. (1954) A decision-making theory of visual
detection. Psychological Review 61:401-09. |DL]
Teghtsoonian, R. (1971) On the exponents in Stevens' Law and the constant in
Ekman's Law. Psychological Review 78:71-80. [aDJM, ANL, MO, RT]
(1973) Range effects in psychophysical scaling and a revision of Stevens'
Law. American Journal of Psychology 86:3-27. [aDJM]
(1974) On facts and theories in psychophysics: Does Ekman's Law exist? In:
Sensation and measurement, ed. H. R. Moskowitz et al. Reidel.
[RT]
Thurstone, L. L. (1927a) Psychophysical analysis. American Journal of
Psychology 38:368-89. [aDJM, DL]
(1927b) A law of comparative judgment. Psychological Review 34:273-
86. [aDJM, DL]
(1927c) A mental unit of measurement. Psychological Review 34:415
23. [aDJM]
(1927d) Three psychophysical laws. Psychological Review 34:424-
32. [aDJM]
(1929) Theory of attitude measurement. Psychological Review 36:222-
41. [DL]
Thurstone, L. L. & Chave, E. S. (1929) The measurement of attitude.
University of Chicago Press. [DL]
Titchcner, E. B. (1896/1911) A text-book of psychology. Macmillan. [MW]
(1901-1905) Experimental psychology: A manual of laboratory practice,
vols. 1 and 2. Macmillan. [aDJM, SL]
Tolhurst, D. J., Movshon, J. A. & Dean, A. F. (1983) The statistical reliability
of signals in single neurons in cat and monkey visual cortex. Vision
Research 23:775-85. [aDJM]
Torgerson, W. S. (1961) Distances and ratios in psychophysical scaling. Ada
Psychologica 19:201-5. [aDJM)
Townsend, J. T. & Ashby, F, G. (1983) Stochastic modeling of elementary
psychological processes. Cambridge University Press
(Cambridge). [rDJM, RDL]
Treisman, M. (1961) The nature of the psychophysical law and its significance
for the scaling of sensory magnitudes. Ada Psychologica 19:213-
14. [MTI
(1962) Psychological explanation: The "private data" hypothesis. Rritish
Journal for the Philosophy of Science 13:130-43. [MT]
(1963a) Laws of sensory magnitude. Nature 198:914-15. [MT]
(1963b) Neural effects and the psychophysical law. Nature 200:1241-
42. [MT]
(1964a) Sensory scaling and the psychophysical law. Quarterly Journal of
Experimental Psychology 16:11-22. [aDJM, MT]
(1964b) The effect of one stimulus on the threshold for another: An
application of signal detectability theory. British Journal of Statistical
Psychology 17:15-35. [aDJM]
(1964c) What do sensory scales measure? Quarterly Journal of Experimental
Psychology 16:387-91. [MT]
(1985) The magical number seven and some other features of category
scaling: Properties of a model for absolute judgment. Journal of
Mathematical Psychology 29:175-230. [MT]
(1989) Sensory scaling: Unanswered questions. Behavioral and Brain
Sciences 12:293-94. [MT]
Turner, R. S. (1982) Helmholtz, sensory physiology, and the disciplinary
development of Cerman psychology. In: The problematic science:
Psychology in nineteenth century thought, ed. W. R. Woodward & M. G.
Ash. Praeger. [rDJM]
Urban, F. M. (1910) The method of constant stimuli and its generalizations.
Psychological Reviciu 17:229-59. [aDJM]
Van Brakel (1990) Units of measurement: Some Kripkean considerations.
Erkcnntnis 33:297-317. [JVB[
Van Brakel, J. & van der Peut, A. A. (1979) Alternative physical concatenation
procedures in length measurements. Abstrads of the 6th International
Congress of Logic, Methodology and Philosophy of Science 8:36-
40. [JVB]
VVaismann, F. (1930-31) Logische Analyse des Wahrscheinlichkeitsbegriffs.
Erkcnntnis 1:228-48. [MH]
Wallach, H. (1948) Brightness constancy and the nature of achromatic colors.
Journal of Experimental Psychology 38:310-24. [aDJM]
(1963) The perception of neutral colors. Scientific American 208:107-
16. [aDJM]
References/Murray: History of psychophysics
Ward, J. (1876) An attempt to interpret Fechner's Law. Mind (O.S.) 1:452-
66. [DAS]
(1904/1927) Philosophical orientations and scientific standpoints. Essays in
Philosophy by James Ward, ed. W. R. Sorley & G. F. Stout. Books for
Libraries Press. (Reprinted 1968.) [DAS]
Ward, L. M. (1972) Category judgments of loudness in the absence of an
experimenter-induced identification function: Segmental effects and
power-function fit. Journal of Experimental Psychology 94:179-
84. [aDJM]
Warren, R. M. (1981) Measurement of sensory intensity. Behavioral and
Brain Sciences 4:175-223. [aDJM]
Warren, R. M. & Poulton, E. C. (1960) Basis for lightness-judgments of grays.
American Journal of Psychology 73:380-87. [aDJM]
(1966) Lightness of grays: Effects of background reflectance. Perception and
Psychophysics 1:145-48. [aDJM]
Warren, R. M. & Warren, R. P. (1956) Effect of the relative volume of
standard and comparison object on half-heaviness judgments. American
Journal of Psychology 69:640-43. [HER]
Wasserman, G. S. (1991) Neural and behavioral assessment of sensory quality.
Behavioral and Brain Sciences 14:192-93. [LEK]
Wasserman, G. S. & Wang-Bennett, L. T. (1989) Unity and diversity of
neurelectric and psychophysical functions: The invariance question.
Behavioral and Brain Sciences 12:297-98. [aDJM]
Watson, A. B. (1979) Probability summation over time. Vision Research
19:515-22. [DL]
(1986) Temporal sensitivity. In: Handbook of perception and human
performance, vol. 1, ed. K. R. Boff, L. Kaufman & J. P. Thomas.
Wiley. [aDJM]
Watson, J. B. (1930) Behaviorism (revised edition). W. W. Norton. [rDJM]
Weber, E. H. (1834) De pulsu, resorptione, auditu et tactu Annotationes
anatomicae et physiologicae. Leipzig: Koehler. [HG]
(1846/1978) Der Tastsinn und das Gemeingefiihl. In: . H. Weber: The
sense of touch, ed. H. E. Ross & D. J. Murray. (Translated by D. J.
Murray.) Academic Press. [HG, SL]
(1851) Die Lehre vom Tastsinne und Cemeingefuhle aufVersuche gegriindet,
fur Aerzte und Philosophen. Braunschweig: Vieweg. [HG]
(1852) Uber den Raumsinn und die Empfindungskrcise in der Haut und im
Auge. Berichte der Koniglichen Sachsischen Akademie der
Wissenschaften. Mathematisch-physikalische-Classe 85-164. [O-JG]
(1978) The sense of touch {De Tactu [1834] translated by H, E. Ross; Der
Tastsinn [1846] translated by D. J. Murray). Academic Press. [aDJM,
HG]
Weiss, D. J. (1975) Quantifying private events: A functional measurement
analysis of equisection. Perception and Psychophysics 17:351-57. [DJW]
(1981) The impossible dream of Fechner and Stevens. Perception 10:431-
34. [LEK, DJW]
(1989) Psychophysics and metaphysics. Behavioral and Brain Sciences
12:298-99. [LEK, DJW]
Welford, A. T., ed. (1980) Reaction times. Academic Press. [rDJM]
Wendt, D. (1982) On S. S. Stevens' psychophysics and the measurement of
subjective probability and utility. In: Social attitudes and psychophysical
measurement, ed. B. Wegener. Erlbaum. [MH]
Wemer, G. & Mountcastle, V. B. (1965) Neural activity in mechanoreceptive
cutaneous afferents: Stimulus-response relations, Weber functions, and
information transmission. Journal of Neurophysiology 28:359-
97. [rDJM, WLB]
Wertheimer, M. (1953) An investigation of the "randomness" of threshold
measurements. Journal of Experimental Psychology 45:294-303. [aDJM]
West, B. J. (1985) An essay on the importance of being nonlinear. Lecture
notes in biomathematics, 62. Springer-Verlag. [RAMG]
Whittle, P. (1991) Sensory and perceptual processes in seeing brightness and
lightness. In: From pigments to perception, ed. A. Valberg & B. B. Lee.
Plenum. [PW]
(1992) Brightness, discriminability and the "crispening effect." Vision
Research 32:1493-1507. [rDJM, PW]
Wiggins, S. (1990) Introduction to applied nonlinear dynamical systems and
chaos. Springer-Verlag. [RAMG]
Wilson, M. (1987) Brain mechanisms in categorical perception. In: Categorical
perception, ed. S. Harnad. Cambridge University Press. [aDJM]
Wolff, C. (1725) Verniinftige Cedancken vom Cebrauche Der Theile in
Menschen, Thieren und Pflantzen, den Liebhabern der Wahrheit
mitgetheilet. Frankfurt/Leipzig: Renger. (Reprinted in [1980] G. 01ms.
Gesammelte Werke I. Abt.: Deutsche Schriften.) [O-JG]
(1731) Psychologia empirica. Reprinted in: Christian Wolff: Gesammelte
Werke, Part II, vol. 5., ed. J. Ecole. Georg 01ms (1968). [rDJM]
(1734) Psychologia rationalis. Reprinted in: Christian Wolff: Gesammelte
Werke, Part II, vol. 6., ed. J. Ecole. Georg 01ms (1972). [rDJM]
BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1 185
References/Murray: History of psychophysics

Woodworth, R. S. & Schlosberg, H. (1954) Experimental psychology.
Methuen. [aDJM]
Wundt, W. (1862) Beitrdge zur Theorie der Sinneswahrnehmung. Leipzig:
C. F. Winter. (The introduction, titled "On the methods of psychology,"
is translated in: Classics in psychology ed. T. Shipley. Philosophical
Library 1961.) [rDJM]
(1874) Cnindziige der physiologischen Psychologie, 1st ed. W.
Engelmann. [rDJM]
Wyszecki, G. (1986) Color appearance. In: Handbook of perception and
human performance, vol. 1, ed. K. R. Boff, L. Kaufman & J. R. Thomas.
Wiley. [aDJM]
Yilmaz, H. (1967) Perceptual invariance and the psychophysical law.
Perception and Psychophysics 2:533-38. [arDJM]
Zwislocki, J. J. & Goodman, D. A. (1980) Absolute scaling of sensory
magnitudes: A validation. Perception and Psychophysics 28:28-
38. [aDJM]

PSYCHOLOGISTS FOR THE ETHICAL TREATMENT OF ANIMALS

Human Innovations and Alternatives, an annual publication

featuring brief articles from various laboratory disciplines that
describe in concrete terms "alternatives" in husbandry, toxicology,
experimental design and procedure and teaching. Individual
subscriptions are $20, institutional $25. For information on
manuscript submission, write to: Emmanuel Bernstein, Ph.D.,
Adirondack Counseling, Glenwood Estate, Saranac Lakes, NY
12983.

Society and Animals publishes studies which describe and analyze

our experience of nonhuman animals. The goal of the journal is
to stimulate and support an emerging content area within the
social sciences consisting of studies of the ways in which
nonhuman animals figure in our lives. Individual subscriptions
are $30 per volume (2 issues). For information on manuscript
submission, write to: Kenneth Shapiro, Ph.D., PSYeta, P.O. Box
1297, Washington Grove, MD 20880-1297.

Subscriptions are payable to: PSYeta, P.O. Box 1297, Washington

Grove, MD 20880-1297.

PSYeta is an independent association of psychologists dedicated

to the promotion of animal welfare within the science and
profession of psychology and within the community at large.

PSY1 eta
PSYchologistsforthe
Ethical Treatment of Animals

186 BEHAVIORAL AND BRAIN SCIENCES (1993) 16:1