Documente Academic
Documente Profesional
Documente Cultură
of Social Dialects
In our attempt to collect maningful data, there are lwo major areas of
concern: (i) the choice of a sample population and (2) the elicitation
o[ adequate speech data.
THE SA]VIPLE
When a sociolinguist decides to describe the speech behavior oi a particular
population, he is l'aced with the problem of defining his univese in such a
way as lo ensure that his observations adequateiy reDresent the population.
The first decision he must make, therefore, is to delirnir the popuiation from
ivhich his sample wiil be drawn. In some cases, it is quite possible to choose
a popuiation that is already defineci in terms of sone arbitrrry boundary'
such as a geographical one. For example, we may decide rhat we are golng
to study the speech of a given locale, suctr as Detroit, tvf ichigan, Rochester,
New York, or Vfeadville, Mississipoi. For a population of this type, the
bouncries oi our universe have aireadlr been es;ablished ibr us, so that our
only task is to select a represenlative smple from hat population. In other
cases, we may '/ant to describe the speech behavior of a populatioo defined
sociaily raher lhan geographicaily, such as a lorver-class inner-city group. If
this is our universe, then we need to estabiish the boundary on the basis of
crireria such as social class, age, and so lorth. Once we have established who
qualifres according o our social criteria, we can decide how o obtain a
representative sample of this population.
In some rypes of socioiogical studies, randoti sampling is used, in rvhich
erch cerson in che iotri popuiation has n equal c:rance oi being selecred
from the sampie- Random sarnpling, horvever, should nor be equated wtth
haphazard or casuai selection. it is an organized procedure for choosrng
che informants in such a wal- as o eiiminate selection bias' One of the
rrditionai wavs of obtarning an authentic rlrndorn sari:ole is rhrough ihe use
Sce:'Fieid \ftho<ls rn the Stud-v oIsocial Dialects'. in Woiiiam, W. nci Fasold'
R. W. (1974) The Stucil, of Sociai Dialects in Amerlcan Engirs iEngiewood Ciitfs' NJ:
Prenrice.Hal[) pp. ]6-;:.
v
,.-o.
informanrs
randomly rvithiri the farnily, nor could I predicr how large an area I would
to cover by this merhod before availabie resources were expendeci.
9: qu,.
in which wc decide !o invesrigate he variables ol'social clss, sex, age, and '
cthnicity. we choose to h4ve a sample represe nting four social classes, both I
sexes, three age groups, and borh black and white subjects. If we wanr each
of the logically possible social categories represented in our sample, we wiil
have a sample distriburjon as in Figure 9.i.
in our samplc, we wanr to make sure thar all cells (e.g. upper middle-class
white l0 to l2-year-old females or lower working-class biack 25 co 35-yearold males) are adequately r.p..s.nt.d, so we chos.'only a given numberof
informants for each. The total number of ceils in the sample is the number
ol categories of each social variable multiplied by each orher.. There are four
sociai classes multiptied by two races, two sexes, and three age groups, and
so the total number of cells is 48,4 x2xzx j:4g. when soeakins of a
r.^!..
nf 'h:
;. ifloi'f
, -:..; ia
-^ :C:lk
r: 15
^-^ -,- Of lhC nUmbel Of infOfmantS
)-iriirL ur
!:r) t"-o
LjL,., ;.
L-.iiai?.-]
-^..^ .in each cell thrn oithe rctal nu;nber of inibrmants. For exarnpie, if we.usr
had i informants in each ol the ce ils in our sample, we would have a sample
of 240 subjects. But il we were conducting a study of sex d.ifferences in the
speech of lcwer-class whites, we would have only one variable. In this case,
a ioral of 100 subjecrs would be divrded equaily inro caregories ol 50males
and 50 females. Thus rhe coral sampie in tfus insrance ,ould be
considerabiy srnaller. but thc represenrarion in each of rhe rwo ceils much
sreater. If rve have a large nurnbe of celis, ir mar be possible io collapse
some of hem for some ryDes ot'analysis, but an anail'sis oi he efrecc of all
the intersecting vanabies on each other will require adecu3te represeniation
oi ail of our
cells.'
lVhl!i
Male
l0-12 l4-17 l5-li
A stnct random
socloungul5Ilc analvsls.
As an aiternaiive ro srici raadorn sarnplrng, i is otien more elicieni to
obtain a represenraive sarnole lor predeterrlineci socral caregories. In rhis
crocedrre. rhe social coroosition oi':he sanpie rs iirst derermined. then
.:i.lt^.'-.-.-
.-':,j'...1..
---t:t . ::
..":,it,.
Femaie
l0-12 l4-i7 l5-li
El:ck
Male
Fenale
l0-ll t.r-i7 ii-15 l0-i2 l4-17 25-li
!{iddle
Class
Lcwer
Viiclc
Class
Clss
Lo
*'e
rilng
1.t;,_.
clie:cnei ln
The question of optim:rl sample size flor the study of social dialects is still
undetermined. On the one hand, there is the tradition of linguistics which
generally relies on very small samples. In some cases, one or just a few
indiduals scrve as informants, and sometimes the linguist acn as his own
informant. On the other hand, thetradition of sociological surveys is to have
rather substantiai numbers o[subjects, often in the'hundreds or thousands.
The investigation of social dialects must rely on samples that are somewhere
. in between these two traditions. It is, however, difficult to even approximate
what might be a reasonable number of informants in each cell. There are
both theoretical and practical considerations. To a certain extent, the size of
the sample is dependent upon how homogeneous behavior can be expected
to be. The more homogeneous the behavior, the smaller the sample may be. It appears that linguistic behavior is considerably more homogeneous than
some other types o.f behavior, so that we might obtain a reliable sample by
using a smaller sample than some types of sociological surveys. Linguists
have a tradition of assuming that the speech of a few informants may be
sufficient to represent a language system, but exactly how many informants
is sufficien in each ceil is still difficult to predict for a given study. in part,
the size depends on the type of sociolinguiscic problem wirh which we are
dealing. If, lor example, we are dealing with subjective reactions to speech
through a formal questionnaire, there is no reason why we cannot use
sampies that are rather large in size. A requirement of l0 to 20 subjects in
each cell would appear to be a minimum in such a case if we expect to arrive
at s.tatisticaiiy significant results. On the other hand, if we are doing a
quantiative analysis of linguistic variables ol the type undertaken by Labov
et'al. (1968), Wolfram (1969, l9i3), or Fasold (t972), the sample must be
limited in size for quite practical reasbns. This type of anaiysis requires
detailed extraction of speech data. The simple procedure of e.xtractioo may
take several days for each informant, so tha! a comprehensive analysis of
hundreds of tapes is precluded if we are to complete our analysis within
a reasonable amount of tirne. The larger the sample, of course, the
rnore social variables we can examine and the more coofidence we can
ha.e in our conclusions, but there are sometimes practical limitations
ot time. It appeers, horvever. tha if we have fewer than five informants in
erch ceil tor this type of analysis. we r!.n the risk of getting quite skerveC
The ne.'< step in approaching the NNE pcer group nd its use of rhe bsic
vernacular was to study a particular recreation center in detail. . . . In the
initial series of discussions and interviews, we locred one major peer
9?
results.
So lar rve h',e relerred only to samples dcoenden on iome type oi'
raniom selection proceCure. But there are studies in sociai dialects fbr which
we ma-v rvrsh io abandon randornness comoierell,' in t'avor of some
struciured pttern of inlbrmant seiection based on other criteria. For
rrmnlp I rhov:d his associates'mosi insishcfui lineuistic,laa on black
adolescent speech in Hariem came not trom his ranJom'sample, bul frqn'
his study oi selected pecr sroups. The selection of pe:r-group inlormans is
descnbed by hrnr as tbilows:
93
group at the Stephen Foster Center, and began to study it using the'S-G-S'
Sociai Strrus
94
community'l There are two main types of Procedures for doing this - those
dependent on thc objective measurement of an outsidc investigator and
those dependent upon the subjective evaluations ol the communily
participants themselves.
)
J
ci
nn
possible
was
is given
in ihe tciiowrng
Level of Edztcation
Any eraCuacc degree lprolessionrl)
At this point,
Skilled workmen
Semi-skilled workmen
Unskilled wokers (Shuy. Wollram, and fule-v 1968: t2)
these give a combined score o[20. For :- oerson rated 7 on both educatton
ancl occuparion (a laborer wilh a third grade ducation' for example)'
wirh a house in a class VI neighborhooi. the score is i3-1. ob'iiously, the
lower lhe number the higher the oresige. end vice versa' iShul"' \\'olfrsrl'
.i.
j
7
It
e Is
2
1
lrit:
Clss
Two types of residency scales are also used, one relating to the individuai
house type and one to the more general dwelling area. with reference to
per
house type, characteristics such as size, condition, number o[ inhabitants
to
overail
room, and plumbing are considered. Dwelling area refers
neighborhood conditions (such as nondeteriorating single homes wrth
,pu-.iout yards cr deteriorating block homes with no yards). Although.this
type of evaluation may appear to be dependenl on the personal impressions
of the .ualu"tors, quite objective measures can be set up on the basis of
Major professionals
Executives of large concernsLesser professionals
Executives of medium-sized concerns
Admirustralors of small businesses
S.m i-nrnlcs
those which de limit actual salary levels (e.g- above $30 000, S20 000-5l0 000'
etc. [at 1974 rates!]) and those which delimit the source of the income
(e.g. inherited income, profits and fees, e tc.). The source of the income is
often more reliable than the actual income levels because income levels can
become outdated quice rapidly and actuai income is not always commensu-'
Occupation
Technicians
Owners of petty businesses
6
lj)
Qcs gverall sccres are ccmputeci' it rs pcssibl: :" dL"idt ihe ce'ouiacion
who had
inro tJiscreie social levels. In rhe Detoit srudl; cited abrlve. subjcis
as ucper middie ciss, those
wirh scces between -19 and 77 lorver middie cilss, thcse with scores between
ll)
96
78 and t0 upper working class, and those belween t07 and l'14 lower
working class. Subjects whose scores fall.at the lower or upper ranges of an
of
if
.,ob
Ralph
ll . Fasold
q'7
referred to
proposition.
informants
on the
basis
ol
techntque'
on lhe
community membrs are asked to evaluate personai acquaintances
to-be
thought
be
my
person
One
basis oicertain imputed character lraits.
of as
thought
be
may
another
whereas
tracks'
the
'tiom the wrong srde of
own
his
'upper crust'. A communiry member may be asked to designate
can
indi'iduais
,o.iul .ut.go.jes and then rank other community'members'
or
aDove
menbers
aiso be asked to rate themselves and other community
of
indication
an
gives
only
not
below therr in the sociai hierarchy. This
how
sho''vs
aiso
bu"
class
sociai
cf
pecple
in
terrns
vie',v
other
how they'
d
rhey view themselves. Ciass is a concept that is generally recognize
would
rank
social
rhrughut American society, and a valid picture of
participanls
upp.oi to lake into account ihe perceptions of ciass from the
cornnunitlrmernbermal,believeihere'areirvesocilc|ssesandanotner,
socrai-ciass
hree. Are rve,justified in assumtng that rhe rnajc'ril'; oornion.of
of questtons
,Jivisions is necessaril-r' rhe authentic one? These efe the kinds
theoretical
a
tha! researchers rvho use an EP procedure musc deel with on
lr>
rnsl
sociological
various social strat. This is the procedure, for example, thar Labov (1966)
adopted in his srudy of social straficarion of English in New york city
and that Shuy, Wolfram, and Riley (1968) followed in their study of
Detroit speech. I.lot only does it provide a more praciical solution to what
may be a very time-consuming problem. but it provides a base for
examining the validity of sociai-class distinctions as reflecred in the
Iinguistic data.
ELICITATIOI\I OF DATA
once we d.ecid.e ivho we want ro interview, lhe nexi step rn research on social
dia.lects is to elicit the dara that we wanr. Ir,lo srudy of social dialects can
hope to succeed without having aderluare data [or anarysis. As we shall see
later, there are a number o[ different types of data and. mehods for
coiiecting this dara ar rhe disposal of he fieldworker. Belore ciiscussing rhese
merl,rods, however, ir is appropriare to consider he erhics of fieldrvork in
socio linguistics.
Fieldwork Ethics
Thcc arc obviously quesrions ol-cthics whcn onc assumes rot. oo\ ''t'r
"
sietcs a purpose in a community. The first one concerns the ethics oi ible- ,
playing itself. ls it deceitful to assume a role which is in conformity to the .,
local role expectations even when these are lar removed from the explicit
purpose of the rescarch? The answer to this question will be decided in
part by the amountof disparity there is between the roie and the purpose.
it will be decided also by the e.\pecrations of the community. It is not
unethical to act in harmony with these expecrations. (1967: 61)
own religious beliett and would have offended the group had
discovered the fraud.
It
they
pleying.
^recnrlcd
memhpr<hin
- i irct
ir nrv
h r-en re
nret:nrt()ls
tell ihe informant we are interesed in. ihis proceCure seems un',larranted.
On the other hand. there is no need co elace all ;he precise de liis of our
"
r00
t0t
interested
in. For
14)
Sponteneous Interriews
The
spontane
curren!
rnalysis call for. Unfortunateiy, the very iacl thec person is being
inerviewed and taoe-recorded is a formidable obstacLe tc obiaining casual
speech. Labov refes to this problem s /,ie observr's partzio.r:'To obrain
the daa most important for linguistic theory.,,ve have to observe horv
nenle
pel
rrhn
rc
nr
h-;-^
i(197?:
lQ-''
I lr
Rv usin-q
rr.
....-- ther,
II ll).
B.u.,
^h"-,,1'
".^sy are. nol being observed'
various techniques, it is possibie ro ncurralize rhe narurai obsiacies inhereot
in anv iniervie,.v situa:ion. The sol of sponieneous incerrrewrng is quice
strarghtlonvard and simple: the inter'",iewer wxnrs to get as rnuch ltee
converstion as poss;ble. i{e r.vnrs i5e rnt',rmni ro lbcus,)n:nc t'Jni oahls
conversation so rhat he pa.,'s minini attenrion to ihe lvav he is sceakrng.
The less attencion paid to his speech, the more iniornei and natura] we crn
expect his speech to be. I! most cases, this means ihai rvhat inibrmarts talk
about is less important than the lac thar the;, i.Ik. Therel'ore, lengthy'
narralives are toierated. and in fact encouraged, e.;en inough ihel malr be
tangential co lhc specilrc cuesrions asked.
t0?
oJ"
l{alt
Sociul Diulect's
tn
obtaining adequate free conversation from these individuals" For the most
part, they constitute a small minority of interviews.
A spontaneous conversation calls for considerable flexibility in what the
informant talks about, but this should not be interpreted to meen lhat no
general outline is followed flor eiiciting'conversation. Certain types of
predict beforehand) lend to
{uestions (not always the ones we might
than
others. It is possible to
readiiy
naturally eiici conversation more
way
as
to get specific types of
a
in
such
interview
an
informaily direcr
the inlormant's peer
things
as
abour.
such
informarion
sociotogiial
associations, social Status, and general patterning of social interacttons, as
well as to obtain a natural representation of certain types of iinguistic
structures. Suppose we wanted to ensure that there was an adequate
representation oi past-tense constructions in our interview. This couid be
accomplished by rsking he intormant to reiate an incidenr from the past.
oil.the other hand, if we wanted to ge! a repfesentation of presenl-tense
forms, we would have to make sufe that we had adequate conversa[lon
about present-time activities. I our analysis of the use of invariant e in
Black ingtish, we observed that its occurrence was oflen concentrated in
srories about the way in which chiidren's games were piayed (e.g- ll/e run
and hide and the last person thLt get to base, they be ir). This observatton was
a cue io the type of convesations from which we mighr hope to eiicic its
elicrt
usage. our informal direccion of spontaneous conversation. then. can
boch sociological and linguistic data.
#;;
playing it.'
anl
rd:\
would make
or 'Drd you play kick the can?' Rather' we
or'Tell
';;;
lYalJium
;"t;J
tit
sets of
including a persons
of the successful interviewer' The
Flexibility in this regard- if'" tu'-'ol"rk
that we have a
to *-hith informants can easily relate assumes
An
of it" to*'nunity in which we are interviewing'
the
o[
or rural area must be aware
interviewer in an inner-ciiy, sttUo'Oan'
differenttypes'oIindigenousactivitiesandinterestsoftherespectivegroups.
descriptions of
"..'"iqtt,i"o,
certain pre-knowledge
For example, in
ion.,-t'*
yo'k'
'o*t
of ou most elegant
training
activities
of the community is
J.
be
Cues of informants interests should
io1 bV
community can be compensated'
lack oi knowtbdge concerning the
interuew'
informant as expressed in the
sensitivity io the interests oithe
t"t
to pick.up tut' tontt*iit-:lojt::t^
abie
Alert fieldwort.., stoulJ be
the
cite
fulev
and
examite' Shuy' Wolfram'
discussion in he intervie* For
of aieitness to lhe interests of the
un-t*"rnpi"
following interchange
"
informant:
Info.irr"nr,
Yes,
difrerent ones
cailec?
qutre
some adolescenrs can be
Skiil in marbie-shooting is'something
3 matler that a goocl marble-shoot:l:::tO
proud of. and naiurail,y *out Lt
lo
Fieidworker: Did
11968:113)
of speaking.
The style of speech by the iriterviewer can also help direct the focus away
from the speech itself. Interviewers should use a casual style in heir own
speech repertoire. This does nor mean thar they should talk exacrl;r like he
informant or talk in a manner in which they would not normally ialk - this
can appear pretentious and insuhing ro the informant. It wou.ld be
pretentious for a white middle-class interviewer to ry to use Vernacular
Black En_eiish when intervierving a black inforrnanr, bur if he can narurally
adopt a nonstandard variety of whire speech, ir may heip in serring an
informal atmosphere for the interview.
j.
t05
104
course o ihe inrerview. The initial questions are generally quite innocuous,
TV,
invotving such topics as chilcihood games, leisure-time activity. movies,
and the like.
"^*.i"ielv
A.
games'
deciding who
the
Do 1,cu tollow any of lhe NY sporis leels? \\'ht Co y'ou ihink of
(or
Namath
Joe
\leti this 1'ear'l Hoi'r abour the Knrcks fcr next ';e3r
r06
Peer Group
How about thc guys you hang around with'? In this group is there one
guy that everybody listens to? How come?
What makes for a leader in the group (tough, hip with girls, good
sounder, etc.)?
Do the guys iri the group sound on each other? How does this work?
What do you sound on? Can it be true, etc.? (If rapport right, ge t some
sounds.)
What makes a good souuder?
Say a new kid moves into the tenement. Any way he can get into your
group?
Who are some of the guys you're tight with? Name some.
Of the guys you named, are there any Negroes? Puerto Ricans in the
group? How about Whites?
Any of these guys speak Spanish? How about their parents?
Aspirations
C.
How about when you're through with schooi? Any idea of what you
might do? What does a
do?
If
someone came
ln the nnai
cire personaiitics
t:iitt:*
ln the group inrerview, a sel of inlormants is tape-recorrl'9,'" up.
generatly
is
.to ,tne
tion wh each other. The topic for discussion
the fieldworker. Ir is expected rhar narural leaders
iton
o;ri,"",,,"irr..
if he is
ii-. grorp will direct lhe conversation' so that the fieldworker'
in the conversation" The '
Dresent, will ot have to actively participate
is that it is the context most
;;;;;.J advantage of the group interview
of rhe interview
conducive to obtainrng caluai'spe"ch. The constrints
the presence of
or
either from the taPe-recording, tire artificial situation'
also lhe most
s
It
setting'
in
lhis
an outsider are most readily ut"o*t
types of
Certain
themes.
indigenous
of
natural setting for the eliciiation
among
i"JG."""t vJrbal activities (e'g' ritualistic insults and sing:ng
group
a
from
Ut".i inn.r-.iry males), in fact, can only be obtained
situation.
interview
"'ii.
advantage by
g.o"p interview has probably been used to its grearest
Labov and his colleagues (ises)
speech rn
exploratory.face-to-face
Harlem. First, the fielworkers conducted some
Then
ieaders'
peer
the
oi
interviews, includin-e some
.acquarnt"nt:-,1"t
outings' Finally' group .sessrons
macle with peer groups in various social
wereconductedinwhlchmultitrackrecordingsweremade.Vfucho|the
who used his
interviewing was conducled by a parricipant observer
sppro'
acrivity
vrbal
elicrr
ro
[""*r.g.ii the indigenous communiry
resulted tn
procedure
This
members'
among
p.., lr,,.o.iioos
;"t"",
some
.A,s
Fasold
be
Group [ntervievs. Probabiy the cioses; \r'e c3':1 come to setling comptetq,f.u-'
naturai speech .in an interview siuatton is by interviewing groups ol peers.
,i*.ittuf
In
wiijo|tenresuliindata|hatisunusabletorhecietailedanlvsistharis
analvsis- Phonological
necessary lor some types of so'crolinguisiic
1::t]:'' 'n
;;;;;;.,.r.
on the tape if
iienrifviag':arious tpttHl,,]',
members 'h-rnsei\.es m"'ha": i'nc:rr"'rn
rviil clominaie gfoup sessions wnli'
must also recognize ihui somt soeakers
io the parr-erns of socil inreracuon
orhers rviil have ver_,- hriie to ,ui,.iu.
adequate iinguislic data may no"
itrui .*ir, in rhe group. For some speakers'
that ihis will hve-to be compensatec
be derived from the eroup session' so
for in iacer iniividuri scssions'
ii*,iry
r08
ot.
(th)
(dh)
casuaJ
for
concentrating
pho no tog ic
fve
ar iab les
..
sg-.
(r)
how
r09
I cin
remember
everything he did: lhis thing, thar thing, and the orher hing. He
used i6 carry three-ne*Tpapers in his mouth ar thamT dme. I
suppos it's thlsame thing with most of uJ-your first dog is like
your first girl. She's more trouble than she's worth, but you can't
seem to forget her.
(Labov 1966:'597)
As illustrated
in this
it
.
'
Finailv. ir is possible to give simple word lisrs thr focus on some of tfre
ciucial phonoiogical realizations of items when one is mosi aware ot.sp*ch.
Anoher type-of word lis is the min.imal pair list, in which trvo items are
read and che inlormant then decides whether these items sound the same or
not. In many cases, the words are distinct in one dialect but poientrally
homophonous in anoher diaiecc (i.e. inerdiaiectai homophony). As
illusiation, w'e can consider the following list used in our studi oi Puerto
fucan Enslish in Hariem:
\ll,rtc!
| tfroronl
,vfdf
rows
roSe
eide
run
hut
rum
shoe
sod
cn!!'
mask
hot
iIIdJJ
soul
deaf
deiih
vo te
eilow
jellcr
srng
tirne
Torn
reign
pln
watch
boil
was h
sold
Doai
sin
raln
west
bel
Wes
ba
pen
bali
r\!'ifi*
l i.
l9-l:l-li)
il0
Ilalt
the above minimul word-pirir list, thirc arc somc itcms th:rt are
homophonous in all dialecrs of English and therelbre nondiagnostic. ltems
like rows and rose a^d rain arld reign represent this category. These items
are inctuded in order to ensure tha the informant understands the task
In
and is responding accordingly. There are a:lso some items we wouid expect
to"be homophonous in certain varielies of Spanish-influenced English,
including bet arrd bat and shoe and clew- These items give us an indication
of rhe extent of Spanish influence on English in formal style. There are also
items we would expect to be homophonous on the basis of the surrounding
Black English dialect, such as deaf and death and pin and pen- Two main
aspects of minimal word lists must be recorded: (l) whether the informant
actually pronounces the items the same or differently and (2) whether he
says that they sound alike or not. The first qualification gives us objective
data while the second gives us an intuitive judgment about the speaker's
sound system. Although intuitive reactions inay appear to be quite
important for an analysis of a phonologicai system, researchers cannot
always take the informant's reac[ioos at flace value. In Some cases there
may be stared differences even though careful analysis (inctuding acoustic
analysis) shows the words to be produced simiiarly, while in other ccses the
converse may be rl.ue.
The E[citation of Subjective Reacfions
spegh data
speech."Yet
questions allowed the informant to exprcss his rerction toward New York
speech and e.xplain why he felt the way he did. Alrhough open-ended
questions concerning subjective reactions allow us to gather data we might
not anticipate in a nondiscursive questionnaire, it can become difficult to
taxonomize and quantify results on this basis. For this reason, many
dull
drcult
poSl tlve
ro ugh
rveak
l.rs t
ilrnnv
"'"r:-
thick
bad
smaf!
vaiuble
e3s!'
neqative
smoolh
strong
slorv
cereiui
simple
ihin
good
dumb
11
more heavily
The closer one checks to one of the poles in the scale' the
A particular
e:(treme.
weighred is his response in terms of the particulaf
results' For
the
of
computation
nurierical value is assigned on this basis for
+l' '
value.of
a
have
may
scale
each
.-r-*Jt, t,".tirig with as the midpoiot,
adjectiva'l^
the
of
side
posidve
the
to
contiguous
+2, and +3 (immediately
side of thel
pole), or -1, -z,ana -: iimmetately cootiglous to tire negative
found
havs
pof*i. Uting the techniiue of lactor analysis, researchers
main
three
into
cluster
inJicating that the judgments tend to
'
.rrirn..
of
evaluation
positiu._neg"tivecontinuumralherthanbymeansofbinarychoice.Used
at
p."p*fy, te semaritic differential can be a useful tool for getng
itemsshowanygreaterrelationshipt,opertlnentlanguagebeha.ioccriteria
use and
irr.o o more'tiaditionai dispositional or role ptaying language
of
types
the traditional
language attitude items'(1969:5). In addition to
tena
included
atrudinal responses asked on his questionnaire, Fishman
irern'commirrnentscaleinwhichuptttoo'twiilingnessorcommitmentlorvas
to language
respond or perform a parricular lype ofactivity with respect
to rnasure
calculated
were
asked
Fishman
questions
*.uru... ih. typ. oi
Spaoish on a
use
of
lhe
strengthen
and
maintain
to
willingness
a person's
ro Partlclpat in
"
to contribure
improling the i"rsoo;s command of Spanish to willingness
the use
uD
.on.V retp d"un.. the activities of an association for building
correlated
were
questions
commiiment
When
the
York'
X.*
oi Spoirit in
significant
a
scale'
wirh the previously given noncomrniirnenr attitude
are
technique
difi'erence arose. Commiimen measures as a ciala-gathenng
can rnore
ihev
because
questionneires
more usetul han traditional ertitude
or
cogrutlve
simple
eliciting
than
rather
rendencies
get
behavioral
ar
direcrly
r'alurtive responscj
made on ihe basis of
So ia, we have oniy referreci to subjecrive reactions
But we can also use
a fieidwoke's questioning or a wntten questionnaire'
to speech' The
reactions
o,f',., ,yp., of stimuii ii eliciting subjecdve
is oe'of
speech
of
samples
tapeirecorded
elicitaion of responses trom
ol social
study
the
in
extent
ionsiderable
to
used
he current technrques
Lambert and
ciiaiecls. One such mehod. originally developed by
his
collcaguesacMcciilUniversitylorevluatingpersonalitytrritsot.
bilingials, has been labeled
the
oneofthedisadvantagesisthatitissometimesdifficulttotindspeaKefswno
dialects we miSht
huu. u.quir. nariuetit. control of the various social
require to Produce such a taPe.
of four socioeconomic
Attheconciusionofthemainpassages.anumberofshortportions(one
in order lo determtne
sentence of lrom 3 to 5 ,"tondt were included
For each of the longer
reactions on the basls of much shorter sdmuli'
the speaker' as well as
of
passages, sub.ects rvere asked to icientily the race
interviervs
attitudinal
the relative socioeconomic class, and to make some
tollows:
as
a semantic differentiai. This rvas oresenled
judgmenis on
r )
a. What is the race of this speaker'l Black 1 IoiWhite
speaker'l
rhis
;. il;, ;; ihe educaionaloccuparional levelgraduate traiilnq
( ) l Cotiege graduare usuaill'- with
personnei msn3sef
( ) 2.
( ) i.
()1.
t t
I la
of thc followins
relaxed
tense
correct
informal
thick
incorrect
of
rhe
videotapes of rhe children from the three different eroups ro cre3re a lype
of ethnic gurse. stereoryping was measured by looking at the extenr to which
the sual picrure determined a particuiar reacrion despire the constancy of
the Standard Engiish. N{odern rechnology atlbrds rhe crearive esearcher a
grert deai of variet,r'in designing wa.r-s of iersins out'various dimensions of
subjec:ive reaclions co language differences.
REFERENCES
Pickeiicn. D. rl9il) 'Inhe:-enr Variabiiiri, rnd \':riabL: Rules .
Langucge. T, pp
-j;-ql
J. D. ( I968) An .lcoustic end
F,rna',t
t.ot,r 0t'
Brvcien,
Soco,l Dnlect .lnai;;:is oi p:rceptua!
variables in Listener [dentiftcation ctnd Rating o.i Negro Sqeakers. usoE projecr
No.
7-C-001.
Fasold. R.
w.
Ane!"-sis (trVashingron.
l5
scalcs:
awkward
lormai
thin
'{
pp. 5- | t.
Labov, W. 11966) The Social Stratifcation of English fi New York Crry flVashington,
DC: Centcr for .dpplied Linguistics).
Labov, W., Cohen, P., Robins, C. and Lewis, J. ( l963) A Study of the Non-Standard
Englbh of Negro and Pierto Rican Spakers in New York City. USOE Final
Report, Research Project No. 3286.
Levine, f,. and Crockett, H. J. (1967)'Fnends'lniiuence on Speech,' Sociological
Inquiry,37, pp.109-28
Samarin, W.I. (1967) Field Linguisrict: A Guide to Linguistic Fietd Vr'ork (New York:
Holt, Rinehart & Winston).
Shuy, R. W., Barau, J. C. and Wolfram, W. (1969) Sociolinguisttc Facrors in Speech
Identifcation. MMH Final Reporr, Projecr No. MH 15048-01.
Shuy, R. W., Wolfrem, W. A. and fuley, W. K. (1968) Fteld Techniques in an Urban
Language Stud.u fWashington, DC: Cenrer for Apolied Linguisrics).
Warner, W. L. (1960) Social Class in America (New York: Harper & Row).
Williams, F (t97i)'Some Research Notes on Dialect Atritudes and Stereotypes', rn
Shuy, R. W. end Fasold, R. W (ed.s), Language Attitules; Currnt Trnds and
Prosoects flVashington, DC: Georgetown University Press)"
Williams, F, Whitehead, J. L. and Ntiiler, L. Vf. (197 l) Auitudinal Correlares of
Children's Spe.ech Characterstics. USOE Research Report Project No. 0-0336.
Wollram, W ( 1969) A Sociolinguistic Description of Detrott Negro Speech
(Washington, DC: Center for Applied Linguisrics).
Wolfram. W (1973)'On What Basis Variable Ruls?', in Studies in:Yew ltr''a:ts oJ'
Analyzrng Variation in English flilashington, DC; Georgetown l-iniversity Press).
Wollram, W. in collaboration with Shiels, N'f. and Fesold, R. W. (197t) Overlapping
[nlluence in the English of Sacond-generation Puerto Riccn Teenagers in Est
Har[em. Final Repon, USOE Projecr No 3-70-0033(:03).