Documente Academic
Documente Profesional
Documente Cultură
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=black.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Blackwell Publishing and The London School of Economics and Political Science are collaborating with
JSTOR to digitize, preserve and extend access to The British Journal of Sociology.
http://www.jstor.org
Problems of Involvement and
Detachment
NORBERT ELIAS
?
Old Lady: Are you not prejudiced
mannoronewho {ells
Author: Madame,rarelywill you meet a moreprejudiced
himselfhe keepshis mindmoreopen. But cannotthat be becauseone part
of ourmind,that svhichwe act with,becomesprejudicedthroughexperience,
andstillwekeepanotherpartcompletely opento observeandjudgewith?
Old Lady: Sir, I do not know.
A"thor: Madame,neitherdo I and it may well be that we are talkingnonsense.
Old Lady: That is an odd termand one I did not encounterin my youth.
A"thor: Madame,we apply the term now to describeunsoundness in abstract
conversation, tendencyin speech.
or, indeed,any overmetaphysical
Old Lady: I must learnto use these termscorrectly.
E. Hemingway, Death in the afternoon.
II
The way in which individualmembersof a group experiencewhatever
affectstheirsenses,the meaningwhichit has for them,dependson the standard
formsof dealingwith, and of thinkingand speakingabout, these phenomena
graduallyevolvedin their society. Thus, althoughthe degreeof detachment
sets of properties. And if this is the assumption underlying one's form of discourse, terms like
" involved " and " detached ", as they are used here, must appear as equivocal and vague.
They have been chosen in preference to other perhaps more familiar terms precisely because
they do not fall in line with linguistic usages which are based on the tacit assumption of the
ultimate independence of psychological and social properties of men. They do not suggest as
some current scientific concepts do that there are two separate sets of human functions or attri-
butes, one psychological and one social in character, which communicate with each other only
occasionally during a limited span of time with a definite beginning and a definite end by means
of those one-way connections which we call " causes-and-effects" and then withdraw from each
other until a new causal connection is established again with a definite beginning and a
definite end.
Both these terms express quite clearly that changes in a person's relation with others and
psychological changes are distinct but inseparable phenomena. The same holds good of their
use as expressions referring to men's relation to " objects " in general. They seem preferable
to others which like " subjective" and " objective " suggest a static and unbridgeable divide
between two entities " subject " and " object ". To give a brief and all too simple example of
their meaning in this context: A philosopheronce said, " If Paul speaks of Peter he teUs us more
about Paul than about Peter." One can say, by way of comment, that in speaking of Peter
he is always telling us something about himself as well as about Peter. One would call his
approach " involved" as long as his own characteristics, the characteristics of the perceiver
overshadow those of the perceived. If Paul's propositions begin to tell more about Peter than
about himself the balance begins to turn in favour of detachment.
228 PROBLEMS OF INVOLVEMENT AND DETACHMENT
shown in one's encounterwith natural forces may vary from indindual to
indisridualand from situation to situation, the concepts themselveswhich,
in societies like ours, all individualsuse in thinking, speakingand acting,
conceptslike " lightning", " tree " or " wolf " not less than " electricity",
" organism", " cause-and-effect " or " nature", in the sense in which they
are used to-day, representa relativelyhigh degreeof detachment; so does
the sociallyinducedexperienceof natureas a " landscape" or as " beautiful".
The range of individualvariationsin detachment,in other words,is limited
by the public standardsof detachmentembodiedin modes of thinkingand
speakingaboutnatureand in the widelyinstitutionalizeduse of naturalforces
for humanends. Comparedwsthpreviousages controlof emotionsin experi-
encingnature,as that of natureitself, has grown. Involvementhas lessened,
but it has not disappeared. Even scientific approachesto nature do not
requirethe extinctionof othermoreinvolvedand emotiveformsof approach.
What distinguishesthese from other less detachedapproachesis the manner
in which tendenciestowardsdetachmentand towardsinvolvementbalance
each other and blend.
Like other people,scientistsengagedin the study of natureare, to some
extent, promptedin the pursuitof their task by personalwishesand wants;
they are often enoughinfluencedby specificneedsof the communityto which
they belong. They may wish to foster their own career. They may hope
that the resultsof their inqulrieswill be in line with theoriesthey have enun-
ciated beforeor with the requirementsand ideals of groupswith which they
identify themselves. But these involvements,in the naturalsciences,deter-
mine as a rule nothingmorethan the generaldirectionof inquiries; they are,
in most cases, counter-balanced and checkedby institutionalizedprocedures
which compel scientists, more or less, to detach themselves,for the time
being, from the urgent issues at hand. The immediateproblems,personal
or communal,induce problemsof a differentkind, scientificproblemswhich
are no longerdirectlyrelatedto specificpersonsor groups. The former,more
narrowlytime-bound,often serve merely as a motive force; the latter, the
seientificproblemswhich they may have induced,owe their forrnand their
meaningto the widerand less time-boundcontinuumof theoriesand observa-
tions evolved in this or that problem-areaby generationsof specialists.
Like other human activities scientificinquiriesinto nature embodysets
of values. To say that natural sciences are " non-evaluating" or " value-
free" is a misuseof terms. But the sets of values, the types of evaluations
which play a part in scientifieinquiriesof this type differfrom those which
have as their frame of referencethe interests,the well-beingor sufferingof
oneself or of social units to which one belongs. The aim of these inqwries
is to find the inherentorderof events as it is, independentlynot of any, but
of any particularobserver,and the importance,the relevanee,the value of
what one observesis assessedin accordancewith the place and functionit
appearsto have withln this orderitself.
In the explorationof nature,in short, scientistshave learnedthat any
NORBERT ELIAS 229
direct encroachmentupon their work by short-terminterests or needs of
specificpersonsor groupsis liable to jeopardizethe usefulnesswhich their
workmayhavein the endforthemselvesor fortheirowngroup. Theproblems
which they formulateand, by means of their theories,try to solve, have in
relationto personalor socialproblemsof the day a high degreeof autonomy;
so have the sets of values which they use; their work is not " value-free",
but it is, in contrast to that of many social scientists, protectedby firmly
establishedprofessionalstandardsand other institutionalsafeguardsagalnst
the intrusionof heteronomousevaluations.l Here, the primarytendencyof
man to take the short route from a strongly felt need to a preceptfor its
satisfactionhas becomemore or less subordinateto preceptsand procedures
whichrequirea longerroute. Naturalscientistsseekto findways of satisfying
human needs by means of a detour the detourvia detachment. They set
out to find solutionsfor problemspotentiallyrelevantfor all humanbeings
and all human groups. The questioncharacteristicof men's involvement:
" What does it meanfor me or for us ? " has becomesubordinateto questions
like " What is it ? " or " How are these events connectedwith others? " In
this form, the level of detachmentrepresentedby the scientist'swork has
becomemoreor less institutionalizedas partof a scientifictraditionreproduced
by means of a highly specializedtraining,maintainedby various forms of
social control and socially induced emotional restraints; it has become
embodiedin the conceptualtools, the basic assumptionsfthe methods of
speakingand thinking which scientists use.
Moreover,conceptsand methodsof this type have spread,and are spread-
ing agasn and again, from the workshopsof the specialiststo the general
public. In most industrial societies, impersonaltypes of explanationsof
naturalevents and other conceptsbased on the idea of a relativelyautono-
mous order,of a courseof events independentof any specificgroupof human
observers,are used by people almost as a matter of coursethough most of
them are probablyunawareof the long struggleinvolvedin the elaboration
and diffusionof these forms of thinking.
Yet, here too, in society at large,these moredetachedformsof thinking
represent only orle layer in people's approachesto nature. Other more
involved and emotive forms of thinking about nature have by no means
disappeared.
1 This concept has been introduced here in preference to the distinction between scientific
procedures which are " value-free " and others which are not. It rather confuses the issue if
the term " value ", in its application to sciences, is reserved to those " values " which intrude
upon scientific theories and procedures, as it were, from outside. Not oniy has this narrow
use of the word led to the odd conclusion that it is possible to sever the connection between
the activity of " evaluating " and the " values " which serve as its guide, it has also tended to
limit the use of terms like " value " or " evaluating " in such a way that they seem applicable
only in cases of what is otherwise known as " bias " or " prejudice ''. Yet, even the aim of
finding out the relatednessof data, their inherentorderor, as it is sometimesexpressed, at approxi-
mating to the " truth ", implies that one regards the discovery of this relatedness or of the
" truth " as a " value ". In that sense, every scientific endeavour has moral implications.
Instead of distinguishing between two types of sciences, one of which is " value-free " while the
other is not, one may find it both simpler and more apposite to distinguishin scientificpronounce-
ments betsveen two types of evaluations, one autonomous, the other heteronomous, of which
one or the other may be dominant.
230 PROBLEMS OF INVOLVEMENT AND DETACHMENT
Thus in fallingill one may find one's thoughtsstray again and again to
the question: " Who is to blame for this ? '> The childhoodexperienceof
pain as the outcomeof an attack and perhapsa certainurgeto retaliatemay
assertthemselveseven thoughunderthe pressureof an overgrownconscience
the attack may appearas deserved}so that one may come to feel, rightlyor
wrongly,one has only oneselfto blame for it. And yet one may accept at
the same time the doctor'smore detacheddictum that this illness followed
primarilyfrom a completelyblind biologicalcourseof events and not from
anybody'sintentions,not fromconsciousor unconsciousmotivesof any kind.
Moreinvolvedforms of thinking,in short, continueto form an integral
part of our experienceof nature. But in this area of our experiencethey
have becomeincreasinglyoverlaidand counterbalanced by otherswhichmake
higher demandson men's faculty of looking at themselvesas it were from
outsideand of viewingwhat they call " mine" or " ours" as part systemsof
a largersystem. In their experienceof naturemen have been able, in course
of time, to fotm andto face a pictureof the physicaluniversewhichis emotion-
ally far from satisfactory,which,in fact, seems to becomeless and less so as
scienceadvances,but whichat the sametime agreesbetterwith the cumulative
resultsof systematicobservations. They have learnedto imposeupon them-
selves greaterrestraintin their approachesto naturalevents and in exchange
for the short-termsatisfactionswhich they had to give up they have gained
greaterpowerto controland to manipulatenaturalforcesfor their own ends,
and with it, in this sphere, greater security and other new long-term
satisfactions.
III
Thusin theirpublicapproachesto nature,men have travelleda long way
(and have to travel it again and again as they grow up) from the primary,
the childhoodpatternsof thinking. The road they have travelledis still far
fromclear. But one cansee in broadoutlinesomeof its characteristic patterns
and mechanisms.
When men, instead of using stones as they found them against human
enemies or beasts, with greater restraint of their momentary impulses,
gradually changed towards fashioningstones in advance for their use as
weaponsor tools (as we may assumethey did at some time), when,increasing
theirforesight,they graduallychangedfromgatheringfniits androotstowards
gro+ringplants deliberatelyfor their own use, it impliedthat they themselves
as well as their social life and their naturalsurroundings,that their outlook
as well as their actionschanged. The same can be said of those later stages
in whichchangesin men'sthinkingabout naturebecamemoreand morethe
task of scientificspecialists. Throughoutthese developmentsthe mastery
of men over themselvesas expressedin their mentalattitudestowardsnature
and theirmasteryover naturalforcesby handlingthem,have growntogether.
The level and patterns of detachmentrepresentedby public standardsof
NORBERT ELIAS 23I
IV
Paradoxicallyenough, the steady increasein the capacity of men, both
for a moredetachedapproachto naturalforcesand for controllingthem, and
the gradualaccelerationof this process,have helpedto increasethe difficulties
which men have in extending their control over processesof social change
and over their own feelingsin thinking about them.
Dangers threatening men from non-humanforces have been slowly
decreasing. Not the least importanteffect of a more detachedapproachin
this field has been that of limiting fears, of preventingthem, that is, from
irradiatingwidely beyondwhat can be realisticallyassessedas a threat. The
formerhelplessnessin the face of incomprehensible and unmanageablenatural
forceshas slowly given way to a feelingof confidence,the concomitant,one
232 PROBLEMS OF INVOLVEMENT AND DETACHMENT
might say, of increasingfacilitation,of men's powerto raise, in this sphere,
the generallevel of well-beingand to enlargethe areaof securitythroughthe
applicationof patient and systematicresearch.
But the growthof men'scomprehension of naturalforcesand of the use
made of them for human ends is associatedunth specificchangesin human
relationships; it goes hand in hand with the growinginterdependenceof
growingnumbersof people. The gradualaccelerationin the incrementof
knowledgeand use of non-humanforces,bound up with specificchangesin
humanrelationsas it has helped,in turn,to acceleratethe processof change
iSs
VI
The chance which social scientists have to face and to cope with this
dilemmamight be greaterif it were not for anothercharacteristicof their
situationwhichtends to obscurethe natureof these difficulties. That is the
ascendancygained, over the centuries,by a manner or style of thinking
which has proved highly adequate and successfulin men's dealings with
physicalevents, but which is not alwaysequallyappropriateif used in their
dealingswith others. Oneof the majorreasonsfor the difficultieswith which
men have to contendin theirendeavourto gainmorereliableknowledgeabout
themselves is the uncriticaland often dogmatic applicationof categories
and conceptshighly adequatein relationto problemson the level of matter
and energy to other levels of experienceand amongthem to that of social
phenomena. Not only specific expectationsas to how perceiveddata are
connectedwith each other, specificconceptsof causationor of explanation
formedin this mannerare generalizedand used almostas a matterof course
in inquiriesabout relations of men; this mechanicaldiffusionof models
expressesitself, too, for example,in the widespreadidentificationof " ration-
ality" with the use of categoriesdevelopedmainlyin connectionwith experi-
ences of physicalevents, and in the assumptionthat the use of other forms
of thinking must necessarilyindicate a leaning towards metaphysicsand
irTationality.
The same tendency towards over-generalization shows itself in many
currentideas of what is and what is not scientific. By and large,theoriesof
sciencestill use as their principalmodel the physicalsciencesften not in
their contemporary,but in their classicalform. Aspectsof their procedures
are widely regardedas the most potent and decisivefactor responsiblefor
their achievementsand arsthe essentialcharacteristicof sciencesgenerally.
By abstractingsuch aspects from the actual proceduresand techniquesof
the physicalsciences,one amves at a generalmodel of scientificprocedure
NORBERT ELIAS 239
which is knownas ' the scientificmethod". In name,it representsthe dis-
tinguishingcharacteristicscommon to all scientific, as distinct from non-
scientific,forms of solving problems. In fact, it often constitutesa curious
compoundof featureswhich may be universalwith others characteristicof
the physical sciences only and bound up with the specific nature of their
problems. It resemblesa generalconcept" animal" formedwithoutreference
to the evolutionarydiversityand connectionsof animalspeciesfroma rather
restrictedobservationalfieldso that structuresand functionscommonperhaps
to all animals,as distinct from non-livingthings and from plants, minglein
it with otherscharacteristiconly of certaintypes of animals,of, say, mammals
or of vertebrates.
The assumptionis that in this generalizedform " the scientificmethod"
can be transferredfromthe fieldwhereit originated,fromthe physicalsciences,
to all other fields,to biologicalas well as to social sciences,regardlessof the
differentnature of their problems; and that whereverit is appliedit will
workits magic. Amongsocial scientistsin particularit is not uncommonto
attributedifficultiesand inadequaciesof their work to the fact that they do
not go far enough in copying the method of physical sciences. It is this
strongconcentrationof theirattentionon problemsof " method" whichtends
to obscurefromtheir view the difficultiesthat springfromtheirsituationand
from their own approachesto the problemsthey study.
The superiorachievementand status of the physicalsciencesitself con-
stitutes a highly significantfactor in the situationof those who workin the
field of social sciences. If, as participantsin the life of a turbulentsociety,
they are constantly in dangerof using in their inquiriespreconceivedand
immovablesocial convictionsas the basis for their problemsand theories,
as scientiststhey are in dangerof being dominatedby models derivedfrom
inquiriesinto physicalevents and stampedwith the authorityof the physical
sciences.
The fact itself that people confrontedwith the task of formulatingand
exploringnew sets of problemsmodeltheir conceptsand procedureson those
whichhave provedtheirworthin otherfieldsis in no way surprisingor unique.
It is a recurrentfeaturein the historyof men that new crafts and skills, and
amongthemnewscientificspecialisms,in the earlystagesof theirdevelopment,
continueto rely on older models. Sometime is neededbeforea new group
of specialistscan emancipateitself from the rulingstyle of thinkingand of
acting; and in the course of this processtheir attitude towardsthe older
groups,as in other processesof emancipation,is apt to oscillate: they may
go too far for a while and may go on too long in their uncriticalsubmission
to the authorityand prestigeof the dominantstandards; and then again,
they may go too far in their repudiationand in their denial of the functions
which the older models had or have in the developmentof their own. In
most of these respects the emergenceof the younger social sciences from
underthe wings of the older naturalsciencesfollowsthe usual pattern.
But therecan have been rarelya situationin whichthe gradientbetween
240 PROBLEMS OF INVOLVEMENT AND DETACHMENT
the comparativelyhigh level of detachmentmanifestin the older branches
of knowledgeand the much lower representedby the youngerbrancheswas
equally steep. In the physicalsciences}it is not only the developmentand
use of a specificmethodforthe solutionof problemsandthe testingof theories,
but the framingof problemsand theoriesitself which presupposesa high
standardof detachment. The same method transferredto social sciencesis
not infrequentlyused for the explorationof problemsand theoriesconceived
andstudiedunderthe impactof stronginvolvernents. Hencethe use, in social
sciences,of a methodakin to that evolvedin the physicalsciencesoften gives
to the formerthe appearanceof a highlevel of detachmentor of " objectivity"
whichthose who use this method are in fact lacking. It often serves as a
means of circumventingdifficultieswhich spring from their dilemmawith-
out facing it; in many cases, it creates a facade of detachmentmaskinga
highly involved approach.
As a result,a crucialquestionis oftenregardedas sealedand solvedwhich
in fact is still in abeyance: the questionwhichof the proceduresandtechniques
of the physicalsciencesare commensurateto the task of social sciencesand
which are not. The abstractionfrom these specificproceduresof a general
modelof the scientificmethod,and the claimoftenmadefor it as the supreme
characteristicof researchthat is scientific,have led to the neglect, or even
to the exclusionfrom the field of systematicresearch,of wide problem-areas
which do not lend themselveseasily to an explorationby meansof a method
for which the physical scienceshave providedthe prototype. In order to
be able to use methodsof this kind and to prove themselvesscientificin the
eyes of the world,investigatorsare frequentlyinducedto ask and to answer
relativelyinsignificantquestionsand to leave unansweredothersperhapsof
greater significance. They are inducedto cut their problemsso as to suit
their method. The exclusiveand seeminglyfinal characterof many currelt
statements about the scientificmethod finds expressionin the strangei-lea
that problemswhich do not lend themselvesto investigationsbv means of
a methodmodelledon that of the physicalsciencesare no concernof people
engagedin scientificresearch.
On closer investigation,one will probablyfind that the tendency to
considera highlyformalizedpictureof this one set of sciencesandtheirmethod
as the norm and ideal of scientificinquiriesgenerallyis connectedwith a
specificidea aboutthe aim of sciences. It is, one mightthink, boundup with
the assumptionthat among propositionsof empiricalsciences, as among
those of pure mathematicsand relatedformsof logic, the only relevantdis-
tinction to be made is that betweenpropositionswhich are true and others
whichare false; and that the aim of scientificresearchand of its procedures
is simply and solely that of findingthe " truth ", of sifting trlle from false
statements. However,the goal towardswhich positive sciencesare striving
is not, and by theirvery naturecannotbe, whollyidenticalwth that of fields
like logic and mathematicswhich are collcernedwith the inherentorderof
certaintools of thinkingalone. It certainlyhappensin empiricalinvestiga-
NORBERT ELIAS 24I
VII
Is it possibleto determinewith greaterprecisionand cogencythe limita-
tions of methods of scientific researchmodelled on those of the physical
sciences? Can one, in particular,throw more light on the limits to the
usefulnessof mathematicalor, as this termis perhapstoo widein this context,
of quantifyingmodels and techniquesin empiricalresearches?
At the presentstate of development,the weightand relevanceof quanti-
fying proceduresclearly differsin differentproblem-areas.In some, above
all in the physicalsciences,one can see to-day no limit to the usefulnessof
procedureswhichmake relationsof quantitiesstand for the non-quantitative
aspects of the relations of data; the scope for reducingother properties
to quantitiesand for workingout, on the basis of such a reduction,highly
adequatetheoreticalconstructsappearsto be without bounds.
In other fieldsof researchthe scope for similarreductionsis clearlyvery
much narrower; and theoreticalconstructsbased on such reductionsalone
often prove far less adequate. Have problem-areas whichdo not lend them-
selvesas well as the physicalsciencesto the applicationof quantifyingmethods
of researchcertaingeneralpropertieswhich can accountfor such differences
in the scope and relevance of quantifyingproceduresas instruments of
research?
242 PROBLEMS OF INVOLVEMENT AND DETACHMENT
It is possibleto think that this problemitself can be readilysolved in
terms of quantitiesalone. As one passes from studiesof matter and energy
and its varioustransformations to those of organismsand their development
as species and individualsand again to studies of men as societiesand indi-
viduals(innot quitethe samesenseof the word),accordingto a not uncommon
view, the problemswhichone encountersbecomesmorecomplex; the greater
complexityis often thoughtto followfromthe fact that the numberof inter-
acting parts, factors, variablesor suchlikeincreaseas one moves from the
study of inorganicmatterto those of organismsand of men; and as a result
of this increasein numbers,so the argumentseems to run, measurements
and mathematicaloperationsgenerally,becomemore and more complicated
and diicult. If one acceptsthe idea that it is the aim of scientificinvestiga-
tions everywhereto explainthe behaviourof compositeunits of observation
by meansof measurementsfromthat of their simplerconstituentparts, each
of the variablesaffectingthe behaviourof such a unit would have to be
measuredby itself so as to determinethe quantitativeaspectsof its relations
with others. The greater the number of variables,the greater would be
the numberof measurementsand the morecomplicatedwouldbe the mathe-
matical operationsnecessaryto determinetheir interplay. In the light of
this hypothesisthe demandsmade on the resourcesin manpower,in com-
puting machines,in mathematicaltechniquesand in money and time would
progressivelyincreasefrom one set of sciencesto the other sviththe increase
in the numberof factorsthat has to be taken into account. Moreand more,
these demandswould becomeprohibitiveand researchon quantitativelines
alone would no longer be possible. Accordingto this view, it is for that
reason that one has to resign oneself to the use of less precise and less
satisfactorymethods of investigationin many fields of studies.
In a way, this approachto the observablelimitationsof quantifying
methodsin researchis itself not uncharacteristicof the mannerin whichforms
of thinkingmost serviceablein the explorationof physicaldata becomedis-
tended into what almost representsa generalstyle of thinking. The choice
of a heap of more and more factors or variablesas a model for increasing
complexityis determinedby a generalexpectationwhich is evidentlybased
on experiencesin physicalresearch,but whichtends to assumethe character
of an a priori belief: by the expectationthat problemsof all kinds can be
satisfactorilysolved in terms of quantitiesalone.
However,the area within which this expectationcan be safely used as
a guide to the formulationof problemsand theorieshas very definitelimits.
The propertiesof differentunits of observationcharacteristicof different
disciplinesare not aloneaffectedby the numberof interactingparts,variables,
factcrs or conditions,but also by the mannerin which constituentsof such
units are connected with each other. Perhaps the best way to indicate
briefly this aspect of these differencesis the hypotheticalconstructionof a
model of models which representdifferentframes of referenceof scientific
problemsin a highly generalizedform as compositeunits arrangedaccording
NORBERT ELIAS 243
to the extent of interdependenceof their constituentsor, more generally,
accordingto the degree of organizationwhich they possess.
Arrangedin this manner,this continuumof modelswouldhave one pole
formedby generalmodels of units, such as congeries,agglomerations,heaps
or multitudes,whoseconstituentsare associatedwith each other temporarily
in the loosest possiblemannerand may exist independentlyof each other
without changingtheir characteristicproperties. The other pole would be
formedby generalmodelsof units such as open systems and processeswhich
are highly self-regulatingand autonomous,which consist of a hierarchyof
interlockingpart-systems and part-processesand whose constituents are
interdependentto such an extent that they cannot be isolated from their
unit without radical changesin their propertiesas well as in those of the
unit itself.
Between these two poles would be spaced out intermediarymodels1
graded accordingto the degree of differentiationand integrationof their
constituents.
As one moves along this continuumof modelsfrom paradigmsof loosely
composedto othersof highly organizedunits, as modelsof congeriesstep by
step give way to those of self-regulatingopen systems and processeswith
more and more levels many of the devices developedfor scientificresearch
into units of the first type change, or even lose, their function. In many
cases, from being the principalinstrumentsand techniquesof research,they
become, at the most, auxiliaries.
Less adequate,in that sense, becomes the concept of an independent
variableof a unit of observationwhichis otherwisekept invariantand, with
it, the type of observationand experimentationbasedon the suppositionthat
what one studies is a heap of potentially independentvariablesand their
effects.
Less adequate,too, becomesthe conceptof a scientificlaw as the general
theoreticalmould for particularconnectionsof constituentsof a largerunit.
For it is one of the tacit assumptionsunderlyingboth the conceptionand the
establishmentof a scientificlaw that the phenomenaof which one wishesto
state in the form of a law that the pattern of their connectionis necessary
and unchanging,do not change their propertiesirreversiblyif they are cut
off from other connectionsor from each other. The type of relationship
whose regularitycan be fairly satisfactorilyexpressedin the form of a law
is a relationshipwhich is impermanentthough it has a permanentpattern:
it can start and cease innumerabletimes without affectingthe behaviourof
other constituentsof the largernexus withinwhichit occursor the properties
1 Even in the elementary form in which it is presented here, such a serial model may help
to clarify the confusion that often arises from an all too clear-cut dichotomy between congeries
and systems. Not all frames of reference of physical problems cluster narrowly around the
congeries pole of the model. Not all frames of reference of biological or sociological problems
have their equivalent close to the other pole. They are, in each of these areas of inquiry, more
widely scattered than it is often assumed. And although, in each of these areas, their bulk can
probably be assigned to a specific region of the serial model, frames of referenceof the problems
of diEerent disciplines, projected on this model, frequently overlap.
244 PROBLEMS OF INVOLVEMENT AND DETACHMENT
of the larger nexus itself. Generallass for particularcases, in short, are
instrumentsfor the solutionof problemswhosereferentialframeis conceived
as a congeries.l
The more the frameworkof problemsresemblesin its characteristicsa
highly self-regulatingsystem and process, the greater in other words the
chancethat constituentsare permanentlyconnectedwith each other so that
they are bound to change their propertiesirrevocablyif these connections
are .severed,the morelikely is it that laws assumea subsidiaryrole as tools
of research; the moredoes one requireas the paramountvehiclefor exploring
and presentingregularitiesof part-connections,system and process-models
clearly representativeof the fact that part-eventsare linked to each other
as constituentsof a functioniIlgunit without whi.chthey wouldnot occuror
would not occur in this manner.
Nor do those time-honouredint.ellectualoperationsknown as iIlduction
and deductionretain quite the same characterthroughoutthis continuum
of models. In their classicalformthey are closelylinkedup with intellectual
movementsup and down between discrete and isolated universals,which
may be generalconcepts,laws, propositionsor hypotheses,and an inEnite
multitudeof particularcaseswhichare also conceivedas capableof preserving
their signiEcantcharacteristicsif they are studiedin isolationindependently
of all other contlectiorss.
When models of multitudes become subordirlateto models of highly
organizedsystemsanothertype of researchoperationgains greaterprominence
modifyingto someextentthoseof inductionanddeduction,namelymovements
up and down between models of the whole and those of its parts.
It is difficultto think of any well establishedterrnsexpressingclearly
the differentialqualitiesand the complementary characterof these two opera-
1 In the case of the second law of thermodynamicsan experimental and statistical law has
been interpreted as a statement about qualities possessed by the referential system as a whole
that is by the physical universe. However, if one may use experiencesin other fields as a model
it is not always safe to assume that propertiesobserved as those of constituent parts of a system
are also properties of the system as a whole. Whether or not one is justified, in this case, to
assume that regularitiesobserved in a part-regic)nof a system, in a part-regionof both time and
space, can be interpretedas regularitiesof the whole system only pShysicistsare entitled to judge.
However, these general considerations about laws are hardly affected by this case. In
physics as in other scientific disciplinesthe referentialframeworkof problemsis far from uniform.
Although, in the majority of cases, the units of observation are simply conceitredas heaps, there
are others in which they are envisaged as units endowed with properties approachingto those
of systems. But compared with the models of systems and processes developed in some of the
biological and some of the social sciences those which have been produced in physical sciences
show, on the whole, a relatively high independence of parts and a relatively low degree of
organization.
This may or may not account for the fact that although the status of laws, in the classical
sense of the words, has to some extent declined in the physical sciences with the ascendance of
models which have some of the characteristics of systems, the change does not appear to be
very pronounced. What apparently has become more pronounced is the implied expectation
that the diverse laws discovered in studies of isolated connections will eventually coalesce and
form with each other a comprehensivetheoretical scaffoldiIlgfor the behaviour of the over-all
system as a whole. Perhaps it is not yet quite clear why one should expect that the unconnected
clusters of connections whose regularities one has more or less reliably determined will subse-
quently link up and fall into pattern. To expect that they will do so, at any rateameans assum-
ing that in the end all congeries including that of ene.gy-matter will turn out to be systems
of a kind or aspects and parts of systems.
NORBERT ELIAS 245
tions. PerhapsoIlemight call " analytical ' those steps of researchin which
the theoreticalrepresentationof a system is treated more or less as a back-
groundfromwhichproblemsof constituentpartsstandout as the primeobject
of researchand as a potential testing-groundfor theoreticalrepresentations
of the whole; and one might call " synoptic" (not to say " synthetic")
those steps which are aimed at forminga more coherenttheoreticalrepre-
sentationof a system as a whole as a unifyingframeworkand as a potential
testing-groundfor relativelyunco-ordinated theoreticalrepresentations of con-
stituent parts. But whatever the technical tenns, one can say that the
solution of problemswhose frameworkrepresentsa highly integratedunit
dependsin the long run on the co-ordinationand balancebetweensteps in
both directions.
In the shortrun,synopsismay be in advanceof analysis. Its theoretical
results have in that case, at the worst, the characterof speculations,at the
best, if they are conformableto a largerbody of observationaland theoretical
fragments,that of workinghypothesis. Many of the ideas put forwardby
the pioneeringsociologistsof the nineteenthcentury, preoccupiedas they
were with the processof mankindas a whole, illustratethis stage. Or else
analysis may be in advanceof synopsis. In that case, knowledgeconsists
of a plethoraof observationaland theoreticalfragmentsfor which a more
unifiedtheoreticalframeworkis not yet in sight. A good deal of the work
done by sociologistsduring part of the twentieth century can serve as an
illustrationof that stage. Manyof them, in reactionfromthe morespecula-
tive aspects of the work done by the system-builderswhich precededthem,
becamedistrustfulof any over-all-viewand of the very idea of " systems"
itself; they confinedthemselvesmoreand moreto the explorationof isolated
clustersof problemswhichcouldbe exploredas nearlyas possibleby methods
used by representativesof othersciencesthoughthey themselveslackedwhat
these othersalreadypossessed: a moreunified,morehighlyintegratedsystem
of theoreticalconstructsas a commonframeof referencefor isolatedstudies
of partwonnections.
In the case of units of observationsuch as multitudesand populations
it is an appropriateaim of researchto developtheoreticalmodelsof a com-
posite unit as a whole by treatingit as the sum total of its componentsand
by tracing back its propertiesto those of its parts. But this reductionof
the whole to its parts becomesincreasinglyless appropriateif one moves
svithinthe continuumof modelstowardsmorehighly organizedunits. As the
constituentsof such units lose their identity if their connectionwith othersis
broken off, as they become and remainwhat they are only as functioning
parts of a functioningsystem of a specifictype, or even of an individual
system, the study of temporaryisolatesis useful only if its resultsare again
and again referredback to a model of their system; the propertiesof parts
cannotbe adequatelyascertainedwithoutthe guidanceprovidedby a theoreti-
cal modelof the whole. At an earlystage in the developmentof a particular
field of problemssuch models,like maps of largelyunexploredregions,may
246 PROBLEMS OF INVOLVEMENT AND DETACHMENT
be full of blanks and perhapsfull of errorswhich can be correctedorlly by
furtherinvestigationsof parts. But howevermuch one or the other may
lag behind,studies on the level of the wholesystem and studies on the level
of part-unitsare greatlyimpededif they cannot rely on a measureof corre-
spondenceandco-ordinationwhichallowsscientiststo move the focusof their
observationsand reflectionsfreely from one level to the other.
VIII
The difficultyis that thereareoftenmorethan two levelsto be considered.
Highly structuredsystems and processeshave often parts which are also
systems and processes; and these in turn may have parts which again are
developingsystems though with a smallermeasureof autonomy. In fact,
such systemswithin systems, such processeswithin processesmay consist of
many levels of varyingrelativestrengthand controllingpowerinterlacedand
interlockedwith each other; so that those who are diggingup knowledgeon
one of them stand in need of free channelsof communicatiorlwith otherswho
are workingin the manygalleriesabove and below and, at the same time, of
a clear conceptionof the position and functionsof their own problem-area,
and of their own situation, within the whole system.
In practice, such lines of communicationare often deficient or non-
existent. Problemson differentlevels are frequentlyinvestigatedby different
groups of specialistswho look hardly beyond their particularpitch. Many
of them draw from limited experienceswith problemscharacteristicof one
levei, or merelyof one of its aspects,inferencesfor the solutionof problems
whosefralneof referencecomprisesmany levels or perhapsthe wholesystem.
And if one of these groups,if, as it has in fact happened,specialistsfor the
study of units which representa relativelylow level of organization,such as
physicists,are greatly in advanceof others in the explorationof their level
and the developmentof correspondingtechniques,the unselect imitationof
their modelsand methodsin studies of morehighly organizedunits is likely
to give rise to a welter of misconceivedproblems.
For not only the whole system, but also each of its constituentsystems
may displaypatternsof connectionsand regularitieswhich are differentand
whichcannotbe deducedfromthoseof theirconstituentsystems. Theoretical
models and methods of researchdesignedfor the study of units which are
less differentiatedand integrated,can be, therefore,at best, only partially
appropriateas means of researchinto more highly organizedunits even if
the latter contain the former or homologuesof the former as constituent
parts.
There are many instances of the difficultiesthat can ensue from the
applicationof modelsdesignedfor the study of part-systemsat one level of
organizationto that of systems at anotherlevel or of the paramount-system
as a whole.
Take, for example,the old controversyabout the usefulnessof physical
NORBERT ELIAS 247
systems such as machinesas explanatorymodelsfor biologicalsystems such
as animalsand men. If one adheresto the traditionalway of thinking,one
can usually perceiveonly two possiblesolutionsto the focal problemof this
controversy. Onecan eitheracceptphysicalsystemsof one liind or the other
as complete models for organismsand assume, explicitly or not, that an
organismas a whole is a set of physicalevents on exactly the same level as
physicalevents outside organisms. Or one can adopt vitalistic models and
assumethat specialnon-physicalforcesareat workin organismswhichaccount
for the observabledifferencesbetween living and non-livingsystems.
In order to accept either of these two alternatives,one has to stretch
a good many points. As in other cases in which it is difficult,not simply
to find a solution for a problem,but to think of any possiblemodel for a
solution which would fit the available evidence reasonablywell, it is the
type of availablemodelsratherthan the evidencewhichrequiresre-examina-
tion. The difficultieswith which men have met, at least since the days of
Descartes,in tackling the question whether or not living systems can be
adequatelyexplainedby analogieswith non-livingsystems are closely bound
up with the traditionof thinkingwhich decreesthat the behaviourof whole
units has to be explainedfrom that of their parts. It becomesless difficult
to conceiveof a more fitting model for the solutionof this questionif it is
acceptedthat there are types of problemswhichrequirea differentapproach
problemswhich can be broughtnearersolution only if one is aware that
the units under observationhave propertieswhich cannot be inferredfrom
those of their parts.
Man-mademachines,as we know them, are homologuesnot of all, but
only of some levels in the hierarchicorderof open systems representedeven
by animalsof a simplertype. As each system of a higherordermay have
propertiesdifferentfrom those lower-ordersystems which form its parts and
as animalsrisingin the evolutionaryscale representsystems within systems
on a steadily rising numberof levels, one would expect the behaviourand
characteristicsof organismsto correspondonly partiallyto those of machines
or of chains of chemicalreactions; one would expect organismsto display
characteristicswhichareonly in someregardssimilarto, but in othersdifferent
from, physical systems, and yet to reveal themselvesas nothing but heaps
of physical particles if their many-levelledorgatlizationis destroyed or if
componentparts are studied in isolation.
But one could no longerexpect, in that case, that all problemsof organ-
ismswill be solvedin the endby analogieswith machinesor with otherphysical
systems and that biologicalscienceswill graduallytransformthemselvesinto
physical sciences. In living systems physical processesare patternedand
organizedin a way whichinducesfurtherpatterningand organizingof these
processesnEven if merl should succeedin constructingartefactswith very
much more and much higher levels of organizationand control than those
of any known machine,artefactswhich could build and rebuildtheir own
structurefromless highly organizedmaterials,whichcouldgrowand develop,
248 PROBLEMS OF INVOLVEMENT AND DETACHMENT
feel and reproducethemselves,one wouldhave to apply to their construction
and to their study biologicalas well as physicalcategoriesand models.
In controversiesbetween vitalists and mechanists,both sides take it
more or less for grantedthat the model of explanationaccordingto which
studies in the propertiesof parts are expected to provide the key for the
problemspresentedby those of the whole, is a universalmodel. In fact, it
is a specificand partial model appropria$eonly to the study of UIlitSon a
relatively low level of organization.l
Or take the much discussedquestion of the relationshipbetween the
behaviourof higheranimalsand that of men. Attemptsto explainthe latter
in terms of the formerare not uncommon. Yet, again, one cannot compre-
hend the functioningand structureof systems which embodya higherlevel
of organizationand control alone in terms of others which are less highly
orgariizedeven if the formerare the descendantsof the latter. While men
functionpartly as other animalsdo, as a whole they functionand behavein
a way no other animaldoes.
The change towards greater cortical dominance(to mention only one
a.spectof these differences)providesa usefulillustrationof the way in which
an increasein the controllingand co-ordinatingpowerof a part-systemon a
very high level in the hierarchyof interlockingsystems goes hand in hand
with changes in the equilibriumand the functioningof systems on all
levels and with a transmogrification of the over-allsystem itself. It is to
differencessuch as these that one will have to turn in order to establish
more clearly and more firmly that and why the sciencesof men cannot be
expected to transformthemselves,sooner or later, into a branch of the
biologicalsciences even though resultsof studiesinto aspects of men within
the competencyof the latter form an integral element of the former.
Finally, similarproblemsand similardifficultiescan be found, again on
a differentlevel and in a differentform,in the long drawn-outdisputeabout
the relationshipof " individual."and " society". Again, one seems to be
left with the choice betweentwo equally unsatisfactoryalternatives. How-
ever muchone may try one'shand at somekind of compromise,on the whole,
opinionsare so far arrayedin two more or less irreconcilablecamps. One
can place orleselfnearerthose who think of societiesas heaps or masses of
indindual people and of their propertiesand their development,simply as
the outcomeof individualintentionsand activities; and one can placeoneself
nearer those who think of societies, of social processesin all their various
aspects,more or less as if they existed in some sense outsideand apartfrom
the individualpeople by whom they are formed.
Commonto both sides, again, is a style of thinking,an idea as to how
1 One need hardly say that the same argument holds good with regard to the old dispute
about the relationship of what is traditionally called " body " and " mind ". In this case
too proposals for the solution of the problem on purely physical and on metaphysical lines are
usually representative of the same style of thinking and equally inept. They may be monistic
or dualistic; they may credit the " mind " with qualities of " matter " or " matter" with
qualities of the " mind ", all these propositionstry to account for the whole in termr of its pa^.
NORBERT ELIAS 249
phenomenaought to be explained,which has been found most serviceable
in men'sattemptsto explain,and to gain controlover, physicalevents. But
in this case the impasseis not only due to the uncrsticaltransferof models
of thinkingfromone fieldto another. Attemptsto workout bettertheoretical
models for the relationshipof individualand society suffereven more from
the fact that this relationshiphas become,in our age, one of the focal points,
if not thefocal point, in the clash of value systems,of socialbeliefsand ideals
which divide some of the most powerfulgroupingsof men. In society at
large, the questionwhat the rights and duties of individualsin society ought
to be, or whether the wellbeingof society oughtto be consideredas more
important than that of individuals,and other questionsof this kind, are
evocative of a wide range of practicalissues which are highly controversial.
Answersto such questionsform in many cases the shibbolethby which fol-
lowers of differentsocial and political creeds recognizefriend and foe. As
a result, reinforcedas it constantlyis by tensions and passionsof rivalling
groups, the question as to what the relationshipof individualand society
ought to be tends to mask and to mufflein discussionsand studiesthe other
as to what kind of relationshipit actually is-so much so that the simple
questionof fact often appearsto be almost incomprehensible.And as it so
happensthat this factualquestionis representativeof one of the basicproblems
of the social sciences,the difiicultieswhich stand in the way of any attempt
to distinguishand to detach it clearly from the topical social and political
questionswhich are often expressedin similar terms constitute one of the
majorbarriersto the furtherdevelopmentof the socialsciencesandparticularly
to that of sociology.
What has been said, so far, about othertypes of part-wholerelationships
can be of some help, if not in solung, at least in clarifyingthis problem. In
manyrespectsthe relationshipbetweenmenas individualsandmenas societies
differsfromthese other types. It is quite unique,and not all its featuresfit
entirely in the schema of a part-wholerelationship. At the same time, it
showsmany of its characteristics and presentsmany of the problemsgenerally
associatedwith it.
All societies, as far as one can see, have the generalcharacteristicsof
systemswith sub-systemson severallevels of whichindividuals,as individuals,
form only one. Organizedas groups,individualsform many others. They
formfamilies; and then againon a higherlevel, as groupsof groups,villages
or towns, classesor industrialsystemsand many similarstructureswhichare
interlockedand which may formwith each other an over-allsystem, such as
tribes, city-states, feudal kingdomsor nation-states,with a dynamicpower-
eqiilibriumof its own. This, in turn, may form part of anotherless highly
organized,less well integratedsystem; tribes may form with each other
a federation of tribes; nation-states a balance-of-power-system.In this
hierarchyof interlockingsocial units the largest unit need not be the most
highly integratedand organizedunit; so far in the history of mankindit
neverwas. But whateverformit may take, that system in the hierarchyof
R
250 PROBLEMS OF INVOLVEMENT AND DETACHMENT
systemswhichconstitutesthe highestlevel of integrationand organizedpower
is also the system whichhas the highestcapacityto regulateits own course.
Like other open systems, it can disintegrateif the pressureof tensionsfrom
within or without becomestoo strong. As long as its organizationremains
more or less intact, it has a higher degree of autonomy than any of its
constituents.
And it is the structureand dearelopment of this system which in the
last resortdeterminesthoseof its part-systemsincludingthoseof its individual
members. Differentlevels in this hierarchyof systems, such as individuals
as such or as families or as classes, have a greater or smallermeasureof
autonomy; they may, for example,co-operateor they may fight with each
other. But the scope for autonomousactions varies with the propertiesof
-theparamountsystem as well as with the location of part-unitswithin it;
and so sloes the basic personalitystructureof its individualmembers. For
on the propertiesand the developmentof this system dependthose of the
institutionalizedset of relationshipswhich we call " family"; this, in turn,
inducesthe organizationand integrationof functionsin individualchildren
who as adultswill be calleduponto.carryon, to developandperhapsto change
the institutionsof the paramountsystem which, by means of this and of
other homeostatic devices, is enabled to perpetuateat least some of its
distinguishingcharacteristics.
Thus unique as the relationshipof " individual" and " society" is, it
has this in common with other part-wholerelationshipscharacteristicof
highly organized,self-regulatingsystems that the regularities,the attributes
and the behaviourof systems on differentlevels and above all those of the
paramountsystem itself cannot be describedsimply in terms appropriateto
those of their parts; nor can they be explainedas effects of which their
-constituentsare the cause. And yet they are nothing outside and apart
from these constituents.
Those who approachsocial phenomena,wittingly or unwittingly,as if
societieswere nothingbut heaps of individualpeopleand who try to explain
the formerin terms of the latter cannot conceive of the fact that groups
formedby individuals,like other organizationsof part-units,have properties
of their own which remainunintelligiblefor an obsenrerif his attention is
focused on individual people as such and not, at the same time, on the
structuresand patterns which individualsform with each other.
Those who approachsocial phenomena,wittingly or not, as if these
phenomenaexistedindependentlyof the individualsby whomthey are formed
areusuallyawareof the fact that phenomenaof this kindhave theirirreducible
regularities. But expecting as they have been trained to expect, that the
regularitiesof compositeunits can be deducedfrom those of their parts and
perhapspuzzledby the fact that they cannot deducethe social regularities
whichthey observesimplyand clearlyfromindividualregularities,they tend
to fall into a mannerof speakingand thinking which suggeststhat social
phenomenaexist in some sense independentlyof individualpeople. They
NORBERT ELIAS 25I