h8r h8r: measure: too many metaphors

Chapter Two

TOO MANY METAPHORS

The metaphor of social construction once

had excellent shock value, but now it has become tired. It can still be

liberating suddenly to realize that something is constructed and is not

part of the nature of things, of people, or human society. But construc-

tion analyses run on apace.1

Looking at their many titles makes one wonder what work the phrase

‘‘social construction’’ is doing. Take the entry for L: The Social Con-

struction of Literacy (Cook-Gumperz 1986). The editor begins with an

article of her own with the same title. There is no indication of what

‘‘social construction’’ means, nor any attempt to exemplify it. The book

is about innovative ways of teaching children to read. The children are

often disadvantaged; then they learn to read, both in and out of the Cali-

fornia school system. Now it certainly is possible to think of literacy—

the idea of literacy—as a social construct, with a good many political

overtones (Hacking 1999). But that was not the point of the book at hand.

It undertakes the valuable task of presenting a ‘‘social perspective’’ on

how children learn to read, or don’t. Why talk of social construction?

We fear a case of bandwagon-jumping.

Construction has been trendy. So many types of analyses invoke social

construction that quite distinct objectives get run together. An all-

encompassing constructionist approach has become rather dull—in both

senses of that word, boring and blunted. One of the attractions of ‘‘con-

struction’’ has been the association with radical political attitudes,

stretching from bemused irony and angry unmasking up to reform,

rebellion, and revolution. The use of the word declares what side one

is on.

Sometimes this declaration tends to complacency. Sometimes utter-

36 TOO MANY METAPHORS

ing the very phrase ‘‘social construction’’ seems more like standing up

at a revival meeting than enunciating a thesis or project. Two things are

readily forgotten. One is that a great many social construction discus-

sions are embedded in the conception of a social problem that began, for

American professors, perhaps a century ago. It led in due course to the

journal Social Problems, and a gifted set of sociologists centered in Chi-

cago. The trouble is that social construction has become a part of the

very discourse that it presents itself as trying to undo.2

Secondly, it is astonishingly easy to lose the whole picture while fo-

cusing on a single pixel. Some constructionists wish to declare a kind

of ownership over the context in which a social problem emerged, with

the view that the outrages of times gone by are the same outrages which

determine the present. This antiquarian view exists as a veneration for

the past—though a strange veneration, which its practitioners would be

insulted to hear so described. Such a position may suffer from myopia,

for ‘‘most of what exists it does not see at all, and the little it does see

it sees much too close up and isolated; it cannot relate what it sees to

anything else and it therefore accords everything it sees equal impor-

tance and therefore to each individual thing too great importance’’

(Nietzsche 1874/1983, 74).

PROCESS AND PRODUCT

Most words ending in ‘‘tion’’ are ambiguous between process and prod-

uct, between the way one gets there, and the result. The termination of

the contract: that can mean the process of terminating the contract. It

can also mean the upshot, the product, the end of the contract. The

pattern is not identical for each ‘‘tion’’ word, because each word nuances

the ambiguity in its own way. ‘‘Production’’ itself can mean the process

of producing, or, in other circumstances, the result of producing. Is the

production of a play process or product? What about movies?

As Lewis White Beck (1950, 27) noted long ago, our word ‘‘construc-

tion’’ shares in this ambivalent pattern of ambiguity. Thus we read, in

a travel guide to Japan, that the construction of the Garden of Katsura

Rikyu, the Imperial Villa in Kyoto, was completed by Toshihito in 1620.

This refers to a process that came to an end in 1620. Then we read that

the garden is a remarkably meaningful formal construction which con-

sists of a semiformal pavement combining cut and irregular stones, fol-

lowed by a series of natural stepping stones, called jumping stones,

TOO MANY METAPHORS 37

which contrast with the stylized cut stones of the villa at the end of the

path. In this sense it is the product that is meaningful, a delicate play

between art and nature, that might not even have been intended by

Toshihito.

Construction-as-process takes place in time. Some social construction

books make this plain in a subtitle. Pickering on constructing quarks

(1986) is subtitled A Sociological History of Particle Physics. Danziger

(1990) on constructing the subject is subtitled Historical Origins of Psy-

chological Research.The recourse to history is implied by other phrases,

as in ‘‘The invention of teenage pregnancy’’ (Arney and Bergen 1984), or

‘‘The ‘making’ of teenage pregnancy’’ (Wong 1997). When Latour and

Woolgar (1979) wrote of the construction of a scientific fact, they wrote

a fragment of the history of endocrinology. It is true that Latour pre-

sented himself as an anthropologist, and many others who write about

the sciences present themselves as sociologists. Nevertheless their in-

dividual case studies are histories. The waters may seem a little mud-

died here. Some of the most prominent early social studies of science

came from Edinburgh in the 1970s. The Edinburgh school, as it was

called, identified its work as sociological, and claimed that it was en-

gaged in a scientific study of science. The theoretical positions of leading

figures such as David Bloor and Barry Barnes, updated in Barnes, Bloor,

and Henry (1996), were more the result of philosophy than sociology.

The empirical work done by the school, well represented also by

MacKenzie (1981) or Mulkay (1979), was historical in character. Trained

historians would often write differently about the phenomenon of teen-

age pregnancy than do Wong (a philosopher), or Arney and Bergen (so-

ciologists), but the description and analysis of the process of construc-

tion, in all these cases, are historical in character.

Construction stories are histories, but to insist on only that angle is

to miss the point. Constructionists about X usually hold that X need

not have existed, or need not be at all like it is. Some urge thatX is quite

bad, as it is, and even that we would be much better if Xwere done away

with, or at least radically transformed. X, the product, is the focus of

attention, although, as I have explained, it is usually not X, the thing,

teenage pregnancy, but the idea of X, the idea of teenage pregnancy, and

the matrices in which the idea has life. If we overhear someone say,

‘‘And even I am a social construction,’’ we know that it is the person as

product who is in question, a person who has been constructed by a

social process, that person’s life history.

38 TOO MANY METAPHORS

Process and product are both part of arguments about construction.

The constructionist argues that the product is not inevitable by showing

how it came into being (historical process), and noting the purely con-

tingent historical determinants of that process.

In the next chapter I turn to natural sciences such as physics, chem-

istry, and molecular biology. Social construction provides one arena for

the science wars. Constructionists state that various items from the

natural sciences are social constructs. Many scientists deny that. They

will admit that there is a (social) history of the discovery of the item in

question, say the Second Law of Thermodynamics. Once upon a time

the Second Law had ideological, political, or religious overtones. That

does not matter. ‘‘The Second Law of Thermodynamics is neither an

empirical claim, nor a social construction, nor a consensus by institu-

tionalized experience, but an inexorable law based on the atomic con-

stitution of matter’’ (Perutz 1996, 69). It is a fact about the universe that

we have discovered. The history of its discovery makes no jot of differ-

ence to what it is, was, and always will be.

Disability

In social affairs, as opposed to chemistry or physics, scholars do make

distinguishable claims, some meaning process and some meaning prod-

uct. Take discussions of disability. We read that ‘‘disability as a category

can only be understood within a framework which suggests that it is

culturally produced and socially structured’’ (Oliver 1984, 15).3 The ‘‘it’’

that is ‘‘culturally produced’’ is a product. The cultural production is

process. The ‘‘socially structured’’ is ambiguous. It could mean that the

product is socially structured, in the sense that it has a structure that

exists in a social setting (a structure reminiscent of the synchronic struc-

ture of Parisian structuralism). Or it could mean that the product is

organized by a historical process named social structuring.

Sometimes process is clearly intended. ‘‘I call the interaction of the

biological and the social to create (or prevent) disability the ‘social con-

struction of disability.’ ’’ (Wendell 1993, 22). Now examine this state-

ment: ‘‘The disabled individual is an ideological construction related to

the core ideology of individualism’’ (Asche and Fine 1988, 13). ‘‘The

disabled individual’’ may refer either to a kind of person, almost a sub-

species (as in ‘‘the whale is . . .’’) or to individuals of that kind, particular

TOO MANY METAPHORS 39

disabled individuals. In either case the author refers to a product, the

kind or the individuals.

There is yet another sense of construction, in addition to product and

process. It has the same etymological roots as, and is similar in meaning

to, ‘‘construal.’’ ‘‘Construal’’ originally meant seeing how a sentence is

to be understood on the basis of its component parts. But the word

quickly acquired the sense of interpretation. In the United States, a strict

constructionist is a constitutional expert who argues for a strict con-

strual of the American Constitution, trying to go no further than the

very words written down and agreed upon by the founding fathers.

Harlan Lane is a distinguished deaf rights advocate, and partisan of

American Sign Language as the basis for a viable linguistic community.

He wrote an essay titled ‘‘The Social Construction of Deafness.’’ There

he mentioned two ‘‘constructions of deafness, which are dominant and

compete for people’s destinies’’ (Lane 1975, 12). What he meant was two

ways of understanding deafness, two ways of thinking about deafness,

two ways of construing deafness. One way to construe deafness is to

think of it as a disability. Another way to construe deafness is to think

of it as the basis for the formation of a linguistic minority. Construal,

construction-as-process, and construction-as-product are inevitably in-

tertwined, but to fail to distinguish them is to fall victim to forgotten

etymologies.

IS ‘‘SOCIAL’’ REDUNDANT?

Most items said to be socially constructed could be constructed only

socially, if they are constructed at all. Hence the epithet ‘‘social’’ is usu-

ally unnecessary and should be used sparingly, and only for emphasis or

contrast. Take for example the G entry for my alphabetical list in Chap-

ter 1. Lorber and Farrell’s anthology (1991) is entitled The Social Con-

struction of Gender. I have already sketched a diversity of feminist ap-

proaches to gender. Yet no matter what definition is preferred, the word

is used for distinctions among people that are grounded in cultural prac-

tices, not biology. If gender is, by definition, something essentially so-

cial, and if it is constructed, how could its construction be other than

social? The point seems to become self-evident when we get to titles

like The Social Construction of Social Policy (Samson and Smith 1996).

The emphasis made with the word ‘‘social’’ becomes useful when we

40 TOO MANY METAPHORS

turn to inanimate objects, phenomena, or facts that are usually thought

of as part of nature, existing independent of human society. This is true

for Latour and Woolgar is (1979) book, subtitledThe Social Construction

of Scientific Facts.They described work done in a laboratory whose head

shared a Nobel prize for medicine for discovering the structure of a cer-

tain tripeptide, a hormone called Thyrotropin Releasing Hormone.

What, according to the authors, was socially constructed? The fact, they

answer, that this hormone was such and such a tripeptide. The hormone,

and the new methods for establishing its structure, were thought to be

so important that they earned the Nobel prize. So it was shocking, in

1979, to be told about the social construction of such an impersonal,

presocial, biochemical fact. Yet in their second edition, Latour and Wool-

gar (1986) dropped the word ‘‘social’’ from their subtitle: ‘‘What does it

mean to talk about ‘social’ construction? There is no shame in admitting

that the term no longer has any meaning . . . By demonstrating its per-

vasive applicability, the social study of science has rendered ‘social’ de-

void of any meaning’’ (p. 281). Latour had his own agenda here, increas-

ingly apparent later with the ‘‘hybrid natural/social actants’’ (Latour

1987) and the ‘‘parliament of things’’ (Latour 1993). He holds that the

usual distinction between the natural and the social is a sham. But one

need not agree with his agenda in urging that we drop the ‘‘social,’’ ex-

cept for an occasional emphasis.

Now turn to essentially social entities, states, or conditions—I strive

for sufficiently generic and noncommittal nouns here—such as literacy

or lesbianism. If literacy is constructed, how other than socially? Per-

haps being lesbian is an innate characteristic of some women, but if

lesbianism is constructed, how other than socially? The philosopher-

sociologists of the natural sciences seem to have been ahead of those

who study more humane topics such as lesbianism or literacy. They

banned the adjective ‘‘social’’ from their titles and their texts. Authors

discussing specifically human affairs continued to employ it rather un-

reflectively.

KANT’S HOUSE

It is not always pointless to use the word ‘‘social’’ in connection with

construction. For example, ‘‘social constructionist’’ has come to name

a quite widespread body of tenets, theories, or attitudes. The adjective

‘‘social’’ is part of the name of this body of thought. Thus Donna Har-

TOO MANY METAPHORS 41

away (1991, 184) wrote that ‘‘recent social studies of science and tech-

nology have made available a very strong social constructionist argu-

ment for all forms of knowledge claims, most certainly and especially

scientific ones.’’ She cited Knorr-Cetina and Mulkay (1983), Bijker et al

(1987), and ‘‘especially Latour’’ (1987) on Pasteur. Although Latour

would erase the adjective ‘‘social,’’ it is useful for Haraway to have a

name for the school of constructionism that she takes to be represented

by Latour, Knorr-Cetina, Mulkay, and Bijker. This is because there are

many other schools. All of them, including social constructionism,

seem to derive from Kant.

Kant was the great pioneer of construction. Onora O’Neill’s book

about Kant, Constructions of Reason (1989), is well titled. Kant was

truly radical in his day, but he still worked within the realm of reason,

even if his very own work signaled the end of the Enlightenment. After

his time, the metaphor of construction has served to express many dif-

ferent kinds of radical philosophical theory, not all of them dedicated to

reason. But all agree with Kant in one respect. Construction brings with

it one or another critical idea, be it the criticism of the Critique of Pure

Reason or the cultural criticisms advanced by constructionists of vari-

ous stripes. We have logical constructions, constructivism in mathe-

matics, and, following Kant, numerous strains of constructionism in

ethical theory, including those of John Rawls and Michel Foucault.

Bertrand Russell’s Logical Constructions

‘‘Wherever possible,’’ wrote Russell, ‘‘logical constructions are to be sub-

stituted for inferred entities’’ (Russell 1918, 155). When you infer an

entity, you infer that it exists. Do numbers exist? Do electrons exist?

We infer (thought Russell) that electrons exist from the reliability of

scientific laws involving electrons. Platonists suppose that numbers ex-

ist. Russell urged ontological caution. He did not like us to infer the

existence of things of some kind, unless we could be certain that things

of that kind do exist. Yet he did not want to follow the skeptic Diogenes

to the bathtub, feigning ignorance about everything. We know a good

deal that we express in terms of numbers and electrons. Russell wanted

to be able to state what we do know, without assuming the existence of

such things. That is where the notion of a logical construction comes

in. On the surface, we appear to be talking about things of a certain kind,

but when we analyze more deeply, we are not.

42 TOO MANY METAPHORS

More technically: Let T be a term that, grammatically, is used to refer

to X, either an individual thing, or things of a certain kind. T is shown

to be a name for a logical construction when sentences in which T ap-

pears are, in a systematic way, logically equivalent to sentences in which

T does not occur, and no reference is made to X. Thus although state-

ments using T appear to refer to X, and hence to imply or presuppose

the existence of X, logical analysis obviates the implication.

What is the point? When an inferred entity X is replaced by a logical

construction, statements about Xmay be asserted without implying the

existence of Xs, since the logical form or deep structure of those sen-

tences makes no reference to X. We are allowed to talk about Xs while

being agnostic about the existence of Xs. This is not a ‘‘same-level’’

skepticism which outright denies that we have grounds for thinking that

there areXs. Russell’s analysis shows that the logical form of statements

about X is not what we think. We discover that below the grammatical

surface we were never talking about so-and-sos in the first place. Rus-

sellian analyses do not debunk inferred entities. They show that there

is no commitment to the existence of so-and-sos. But they do license

statements about so-and-sos, precisely because they show that those

statements do not have the existential commitments we expect them

to have.

Logical Positivism

Logical positivism, usually thought of as antagonistic to construction-

ism, was also deeply committed to the construction metaphor. Russell’s

program was energetically pursued in Rudolf Carnap’sDer logische Auf-

bau der Welt (1928). The English translation renders Aufbau as ‘‘Struc-

ture,’’ butAufbaumeans ‘‘construction’’ (or, in context, ‘‘building’’), and

that is what Carnap meant. He wanted to establish that the world could

be built up from elements, the data of sensory experience, or perhaps

items that played a role in physical science.

The logical positivists (aside from Otto Neurath) might have been

troubled by some of the twists of constructionism in recent sociology.

Not too upset: Thomas Kuhn is standardly presented as the originator

of the modern trend toward social studies of science, but as Peter Gali-

son (1990) has shown, there is a good deal in common between Kuhn

and Carnap, and both men knew it. The roots of social constructionism

TOO MANY METAPHORS 43

are in the very logical positivism that so many present-day construc-

tionists profess to detest.

Yet we should not overdo that statement. Kuhn said little about the

social. More than once he insisted that he himself was an internalist

historian of science, concerned with the interplay between ideas, not

the interactions of people. His masterpiece, ever fresh, is now over

thirty-five years old—truly the work of a previous generation.The Struc-

ture of Scientific Revolutions is rightly honored, by those who conduct

social studies of the sciences, as their pre-eminent predecessor. Yet for

all that Kuhn emphasized a disciplinary matrix of one hundred or so

researchers, or the role of exemplars in science teaching, imitation, and

practice, he had virtually nothing to say about social interaction.

Construct Validity in Empirical Psychology

Before turning to a later genre of construct-ism, another of Russell’s

heirs should be mentioned—this one from empirical psychology. In the

late 1930s logical positivist philosophers of the natural sciences had be-

gun using the noun ‘‘construct’’ for theoretical entities such as electrons

(see Beck 1950 for references). It was taken up in fundamental debates

in the philosophy of the social sciences, for example, in connection with

historical individualism, where you find J. W. N. Watkins, a Popperian,

challenged by May Brodbeck, who studied with Herbert Feigl, the dis-

tinguished logical positivist who had emigrated from Berlin to Minne-

sota. Watkins introduced the ‘‘anonymous individual,’’ which Brodbeck

denounced as an irreducible theoretical construct and thereby unworthy

of a scientific sociology. (For a summary of 1950s debates, with refer-

ences, see May 1987, 14–18.)

After World War II this usage was also transferred to the philosophy

of experimental psychology (for example, MacQuorquodale and Meehl

1948). Hypothetical entities or quantities in psychology came to be

called constructs. Familiar examples are IQ, or Spearman’s controversial

g, the factor called ‘‘general intelligence.’’ How can we distinguish con-

structs that logical positivists took to be virtuous from those that they

took to be suspicious, such as libido? When are hypothetical constructs

valid? The most authoritative text on psychological testing states that

‘‘The term ‘construct validity’ was officially introduced into the psy-

chometrist’s lexicon in 1954 in the Technical Recommendations for

44 TOO MANY METAPHORS

Psychological Tests and Diagnostic Techniques, which constituted the

first edition of the 1985 Testing Standards. The first detailed exposition

of construct validity appeared the following year in an article by Cron-

bach and Meehl (1955)’’ (Anastasi 1988, 161).

The logical positivist ancestry of construct validity has been some-

what suppressed in psychology’s self-history. In 1955 Lee Cronbach

(b. 1916) was rapidly establishing himself as a leading figure in educa-

tion. Paul Meehl (b. 1920), one of the most sophisticated critics of much

experimental and statistical psychology, was another associate of Her-

bert Feigl. Russell’s logical constructions and Carnap’s Aufbau were

very much present at the birth of that cardinal concept of psychological

testing, construct validity.

Nelson Goodman’s Constructionalist Orientation

Nelson Goodman, a philosopher of both the arts and the sciences, has

described his philosophical orientation as ‘‘skeptical, nominalist, and

constructionalist’’ (Goodman 1978, 1). ‘‘Constructionalist’’ seems to be

a word of Goodman’s invention. Possibly two meanings are packed into

this label. One refers to Goodman’s early work. It involves making or

exhibiting constructions. Goodman and Quine (1947) published ‘‘Steps

towards a Constructive Nominalism,’’ dedicated to a systematic elimi-

nation of, among other things, classes, in favor of logical constructions.

Goodman’s The Structure of Appearance (1951), based on his doctoral

dissertation (1940/1990), was the heir to Carnap’sAufbau.His early ver-

sion of constructionalism was an active philosophy which constructed,

or showed how to construct.

It was also a critique of Aufbau, arguing that what we call the world

could be constructed in many ways. Might some ways be simpler than

others? No. Goodman is the author of the most trenchant of critiques

of the notion that simplicity has any existence outside of the eye of the

beholder. Any one world may be made in many ways, and many worlds

may be made.

Goodman’s philosophy evolved from Russell and Carnap. His title,

Ways of Worldmaking (1978), means what it says. Goodman content-

edly talks of making worlds, and takes for granted that it is we, people,

who make them. Moreover, we do so in concert. This sounds social, but

Goodman got there in a straight line from Russell and Carnap.

Goodman and his fellow constructionalists say almost nothing about

TOO MANY METAPHORS 45

actual societies or social processes. This is to some extent a generational

effect. Goodman’s collaborator, W. O. Quine, wrote a great deal about

translation, but it tended to dwell on translation involving imaginary

explorers encountering natives who live in jungles populated by fauna

unknown to any real jungle, namely rabbits. Whatever be the case with

Quine, whose philosophy is more given to regimentation than inquiry,

Goodman’s world-making has to be social: it is people who do it. Good-

man has been enthusiastic about at least some social studies of construc-

tion in the natural sciences.4 Yet his work gives no hint of any actual

social process involved in world-making. Chapter 5 below starts to fill

the gap with a single example; many more are needed.

Constructivism in Mathematics

Kant’s house has many mansions. Kant began his Critique of Pure Rea-

son by trying to understand a puzzle about the truths of arithmetic and

geometry. How can we know them just by thinking, and yet apply them

in the real world which exists independently of thought? The answer

comes in two parts. First, all experience is in space and time, which is

not a fact about experience, but a precondition for anything we call ex-

perience. Second, space is structured by the laws of geometry, and the

units of time are structured by the laws of arithmetic. Both structures

derive from the nature of thought itself. Thus the laws of geometry and

arithmetic are a priori, yet anything experienced must conform to them.

Hence the famous doctrine of the synthetic a priori. Kant’s view of ge-

ometry was devastated at the beginning of this century, when it became

clear through Einstein’s use of Riemann’s mathematics that the real

world might best be described by non-Euclidean geometry: there was no

pure geometry of the mind uniquely best suited for experience.

A second, arithmetical, revolution failed to take place. The Principia

Mathematica of Whitehead and Russell (1910) was intended to undo

Kant’s views of arithmetic, showing that number theory could be de-

duced from pure logic—the numbers and all their properties were, rather

literally, logical constructions. That program did not pan out, for very

famous reasons, connected with the name of Kurt Gödel. And at the

very same time that Whitehead and Russell were writing their opus, a

rival program named intuitionism was inaugurated in Holland by L. E. J.

Brouwer. The ‘‘intuition’’ in question had a technical connotation, al-

luding to what Kant called pure intuitions of space and time. According

46 TOO MANY METAPHORS

to Brouwer, number-theoretic knowledge has two sources. The first

source is a rather Kantian pure intuition of number. Numbers are gen-

erated by us, as they structure the experience of counting. The second

source is proof, and all that we can build up from those intuitions by

proof. Proofs are generated by us, as active thinkers. That seems like a

truism, but it was taken seriously by intuitionists, with remarkable con-

sequences.

Intuitionists held that mathematical objects do not exist until they

have been built up by proofs of their existence, that is, until they have

been constructed by mental operations. Valid proofs must be construc-

tive; that implies that a mathematical object can be assumed to exist

only when, by proof, we have been able to construct it out of intuited

entities. Mathematics, so often thought of as a body of eternal truths,

takes place in time, and objects come into being as they are constructed.

This approach has a well-known radical consequence. You cannot as-

sume the law of the excluded middle. You cannot assume that for any

proposition p, either p is true, or not-p is true. That is because the prop-

osition may refer to an object that has not yet been constructed by proof.

The first years of the twentieth century were revolutionary times in-

deed. Einstein had dethroned Kant, while Brouwer’s intuitionistic rea-

soning challenged Aristotle. Next in line were Lenin and the new quan-

tum mechanics, the one trying to undo capitalism and and the other

undoing causality.

Brouwer’s intuitionism led to various types of what are called con-

structive mathematics, especially constructive analysis (calculus)

(Bishop 1967). As with other construct-isms, constructivism in mathe-

matics is skeptical, because it allows us to assert the existence of objects

only after we have constructed them in a sequence of mental operations.

Hence it forbids us to assert the existence of many mathematical objects

that most mathematicians take for granted—the continuum of real

numbers, for example.

Moral Theory

I have said nothing about ethics, nor will I in these chapters. Let us

record, however, that it has been a constant thrust in moral theory, from

Immanuel Kant’s categorical imperative to John Rawls’s theory of jus-

tice and Michel Foucault’s self-improvement, to insist that the demands

of morality are constructed by ourselves, as moral agents, and that only

TOO MANY METAPHORS 47

those we construct are consistent with the freedom that we require as

moral agents. Some readers may find it natural to couple the names of

Rawls and Kant but bizarre to pair Kant and Foucault. On the contrary,

Foucault began his intellectual career with Kant’s Anthropologie.

Georges Canguilhem was on the mark in calling Les Mots et les choses

a study of the historical a priori. Foucault was pursuing, in his own in-

imitable and transformative way, Kantian ethical themes of the well-

made life in his own final days.

Different names for different construct-isms

Kant may have cast the mold, but the drive for construction belongs to

the twentieth century. The constructing attitude is skeptical. It is also

humanist. It says that the demands of morality do not come from the

idealized and not-human Father or even the idealized posthuman Son.

They come from the demands on rationality that free human agents

place on themselves. It says that mathematical objects are not out there

in Plato’s nonhuman heaven; it is we who bring them into being. It says

that we should not infer the existence of minute and unobservable en-

tities from their causal effects; instead we are to describe phenomena as

they appear to us, analyzing the theoretical entities into logical con-

structions. It says that in experimental psychology we do not use cate-

gories found in nature but constructs whose validity is established by

our practices. To cap it all, Nelson Goodman tells us about ways of

world-making. Not even the world itself is safe from these philosophies

of construction. It is chiefly in this company that the adjective ‘‘social’’

marks out a further theme. Social constructionists teach that items we

had thought were inevitable are social products.

What are we to call these different mansions built within Kant’s

house? We can help ourselves to labels that are almost ready-made.

Goodman called himself a constructionalist. So let constructionalism

refer to the philosophical projects of Russell, Carnap, Goodman, Quine,

and their associates and followers. They aim at exhibiting how, or prov-

ing that, various important entities, concepts, worlds, or whatever are

constructed out of other materials.

Constructionalists may hold that constructions are made by people,

together, but they do not study historical or social events or processes.

Their instincts are skeptical about constructed items, and yet not pro-

foundly so. They do not say flatly that the items do not exist, or that we

48 TOO MANY METAPHORS

cannot have grounds for believing they exist. On the contrary, we have

excellent grounds, but after analysis we realize that our beliefs are not

what they seem. Constructionalism is a change in the level of discourse.

I see this attitude as including not only Brodbeck’s critique of Watkins,

but also the Cronbach and Meehl proposals—now so entrenched in the

experimental psychology of measurement—for legitimating constructs

in psychology which do not derive from direct observation.

Without placing any weight on the terminology, I find it convenient

to leave the label constructivism to mathematics. That is where the

term was first used, at least in modern times, and it denotes a flourish-

ing, if minority, research activity. If we left ‘‘constructivism’’ to math-

ematics, we would avoid the confusion invited by a title such as Social

Constructivism as a Philosophy of Mathematics (Ernest 1998), which

suggests, to anyone who knows anything about mathematical construc-

tivism, something like a social variant of Brouwer’s program (a rather

incoherent idea). It would have been better, I think, to speak of social

constructionism as a philosophy of mathematics, a philosophy that

would presumably maintain that in some sense mathematical objects,

such as numbers, and mathematical facts—theorems—are social con-

structs. That would be analogous to constructionism about the natural

sciences, although the arguments would presumably be different.

It is true that many people nowadays speak of social constructivism

rather than constructionism in any context whatsoever. Throughout

Chapter 1 I spoke instead of (social) constructionism. (I suspect that

some readers, out of habit, actually pronounced the word as ‘‘ivism,’’

not as ‘‘ionism.’’) Nothing should hang upon a spelling, or a syllable, but

my usage does pay attention to the fact that recent enthusiasm for social

constructs is only one mansion in Kant’s big house, and it allows the

others, such as mathematics, to keep the names that they chose for

themselves quite a long time ago.

Hence by constructionism (or social constructionism if we need, on

occasion, to emphasize the social) I shall mean various sociological, his-

torical, and philosophical projects that aim at displaying or analyzing

actual, historically situated, social interactions or causal routes that led

to, or were involved in, the coming into being or establishing of some

present entity or fact.

Most constructionists have never heard of constructivism in mathe-

matics. Constructivists, constructionists, and constructionalists live in

different intellectual milieus. Yet the themes and attitudes that char-

TOO MANY METAPHORS 49

acterize these isms are not so different. From all three we hear that

things are not what they seem. All three involve iconoclastic question-

ing of varnished reality, of what the general run of people take for real.

Surprise, surprise! All construct-isms dwell in the dichotomy between

appearance and reality set up by Plato, and given a definitive form by

Kant. Although social constructionists bask in the sun they call post-

modernism, they are really very old-fashioned.

BUILDING, OR ASSEMBLING FROM PARTS

Construction has become stale. It can be freshened up if we insist that

the metaphor retain one element of its literal meaning, that of building,

or assembling from parts. After the plethora of titles cited at the start of

Chapter 1, it is good to be brought back to the real world, with a book

title such as Constructing a Five String Banjo.When it comes to banjos,

we are told how to make one. Most of the (social) construction/con-

structing works do not exhibit anything resembling a construction.Con-

struction has become a dead metaphor. That expression, itself a meta-

phor, is from Fowler’s eccentric Modern English Usage:

METAPHOR. 1. Live & dead m. In all discussion of m. it must be

borne in mind that some metaphors are living, i.e., are offered &

accepted with a consciousness of their nature as substitutes for their

literal equivalents, while others are dead, i.e., have been so often

used that speaker & hearer have ceased to be aware that the words

used are not literal; but the line of distinction between the live &

the dead is a shifting one, the dead being sometimes liable, under

the stimulus of an affinity or a repulsion, to galvanic stirrings in-

distinguishable from life. (Fowler 1926, 348–49.)

If we are to return ‘‘construction’’ to life, we should attend to its ordinary

meanings, as in constructing a five-string banjo. The core idea, from

Latin to now, is that of building, of putting together. The fairly new

(1992)American Heritage Dictionary first offers ‘‘to form by assembling

or combining parts.’’ Then it gives us a dead metaphor, lacking brick

and mortar, or girders and concrete: ‘‘To create (an argument or a sen-

tence, for example) by systematically arranging ideas or terms.’’ This

metaphor, like the very ancient and very dead geometrical metaphor of

constructing with a ruler and compass, retains the sense of systematic

50 TOO MANY METAPHORS

arrangement of elements, which become part of a whole. Of course the

whole is more than the sum of the parts, because it is a systematic

arrangement, a structure. Buildings are always more than the sum of

their parts.

Constructionalists (Russell) and constructivists (Brouwer) were true

to the root metaphor of construction as building. Whitehead and Russell

wrote down the construction of the number 1 and its successors within

their system of logic. Brouwer had well-understood criteria for the build-

ing up of a mathematical object by proof. And although I would not argue

the point here, it seems to me that in ethics, Kant, Rawls, and Foucault,

to repeat the names of the three moralists I have mentioned, tell us how

to build, and why. I urge (social) constructionists to keep the same faith.

Anything worth calling a construction was or is constructed in quite def-

inite stages, where the later stages are built upon, or out of, the product

of earlier stages. Anything worth calling a construction has a history. But

not just any history. It has to be a history of building. There is no harm in

one person stretching a metaphor, but when many do, they kill it.

Most writers never reflect on the metaphor in ‘‘construction.’’ Sergio

Sismondo is the rare philosopher who does. He generously notices six

legitimate metaphorical uses of the word in the social construct litera-

ture. In fact one of these is not a metaphor at all: scientists ‘‘construct,

through material intervention, artifacts in the laboratory’’ (Sismondo

1996, 50) Surely a great deal of apparatus is literally, not metaphorically,

built out of, or assembled, from parts? Sismondo is insightful when he

includes the root philosophical metaphor of construction which, as I

observed above, derives from Kant. In contrast to Sismondo, I would,

however, insist that most social construct writing is almost wildly

metaphorical, or rather, passes beyond metaphor. Rather than give in-

vidious examples, it is better to mention a few authors in whose work

the construction metaphor is put to good use.

The Psychological Subject

Kurt Danziger’sConstructing the Subject (1990) is a fine example of how

the construction metaphor can be used, fairly literally, when applied to

a social rather than a natural science. Danziger has written a history of

experimental psychology. The subject in question is none other than

‘‘the subject’’ who appears in the experimenters’ laboratory report, once

TOO MANY METAPHORS 51

upon a time often abbreviated by the letter ‘‘s’’ to depersonalize the

subject as much as possible. Today we are all subject to such tests and

expect to be given them when we are growing up, are inducted into the

military, try out for a corporate job, or report an inexplicable malaise to

a psychiatrist.

Danziger writes about the social construction of the subject. But what

is that? As is quite common in Constructing books, Danziger writes

about constructing at least four distinct kinds of entity: a concept or

idea, a practice, a body of knowledge, and individuals themselves. First,

there is the idea of the subject to observe or to test in experiments.

Danziger is convincing when he urges that this is not a self-evident idea

that was well understood as soon as the idea of laboratory-style experi-

mentation on the human mind came into being. The first subjects of

psychological experiments were commonly the experimenters them-

selves—Gustav Fechner, for example. Or the experimenter and subject

were two people who took turns switching roles: the subject becoming

the experimenter who subjected the former experimenter to test. This

contrasts dramatically with the subsequent notions of an objective psy-

chology, in which the subject is thought of as an object s that must be

scrupulously set apart from the experimenter in order to avoid contam-

ination.

Secondly, Danziger’s book is about constructing a family of practices

within which the subject is embedded. The upshot is a laboratory that

is expanded to occupy the worlds of business, the military, education,

law, and pathology, where people are regarded as subjects for testing. In

a powerful passage at the end of his book, Danziger writes of ‘‘a funda-

mental convergence between contexts of investigation and contexts of

application’’:

the individuals under investigation became the objects for the exercise

of a certain kind of social power. This was not a personal, let alone

violent, kind of power, but the kind of impersonal power that Foucault

has characterized as being based on ‘‘discipline.’’ It is the kind of power

that is involved in the management of persons through the subjection

of individual action to an imposed analytic framework and cumulative

measures of performance. The quantitative comparison and evaluation

of these evoked individual performances then leads to an ordering of

individuals under statistical norms. (Danziger 1990, 170)

52 TOO MANY METAPHORS

A third item to be constructed is knowledge. (Danziger’s last chapter is

titled ‘‘The Social Construction of Psychological Knowledge.’’) The pas-

sage quoted above continues as follows:

Such procedures are at the same time techniques for disciplining indi-

viduals and the basis of methods for producing a certain kind of knowl-

edge. As disciplinary techniques the relevant practices had arisen dur-

ing the historical transformation of certain social institutions, like

schools, hospitals, military institutions, and, one may add, industrial

and commercial institutions. . . . This kind of knowledge was essen-

tially administratively useful knowledge required to rationalize tech-

niques of social control in certain institutional contexts. Insofar as it had

become devoted to the production of such knowledge, mid-twentieth-

century psychology had been transformed into an administrative sci-

ence.

Only by implication does Danziger discuss a fourth category, individual

people. We are now trained to answer questionnaires or perform various

tasks in order to find out our talents or what ails us. Of course the tests

themselves do not settle things. Some readers will wish I had followed

the advice given after my vocational aptitude tests early in high school—

that I should become a meteorologist. The point is not what the tests

say about each of us, but that each is now a kind of person who hardly

existed a century and a half ago: fit subject for testing. Without us as

common fodder for tests, there could hardly be such a thing as theMen-

tal Measurements Yearbook (Mitchell, 1992). This handbook is scru-

pulous in admitting only very well validated and widely used tests.

(Meehl’s construct validity is strictly enforced!) The number of available

tests has doubled with each edition over the past decades.

Danziger’s book is a paragon of fairly literal constructionism. It pres-

ents a history of crafting various parts that are in turn assembled into

larger structures. Experimental psychology begins with the physiology

laboratory as model. Through the use of that model a new type of in-

vestigation is constructed. Certain types of inquiry are pared away from

it—Wundt’s introspection, for example. A new element is added. Sub-

jects are not treated individually; aggregates become essential as statis-

tical technologies are advanced. Statistical procedures from agronomy

or biometrics are incorporated, often in black-box form; the psycholo-

gists who use the tests often have little idea of their rationale. There

TOO MANY METAPHORS 53

have been meta-experiments in which fully accredited psychologists

were asked what a significance level means; only a minority give meth-

odologically sound answers.

The metaphor of construction fits the chain of events that Danziger

organizes. This is because there is something of a historical step-by-step

building of specific techniques, institutions, and problems, each using

previous steps, and assembled to form a further stage in the production

of later techniques, institutions, and problems.

UNMASKING

Chapter 1 mentioned another metaphor, the metaphor of unmasking. It

goes back to a familiar predecessor of constructionism—exactly contem-

porary with logical positivism. In his definitive 1925 paper on the so-

ciology of knowledge, Karl Mannheim stated the four factors that cre-

ated a need for the sociology of knowledge:

(1) the self-relativization of thought and knowledge,

(2) the appearance of a new form of relativization introduced by the

‘‘unmasking’’ turn of mind,

(3) the emergence of a new system of reference, that of the social sphere,

in respect of which thought could be conceived to be relative, and

(4) the aspiration to make this relativization total, relating not only

thought or idea, but a whole system of ideas, to an underlying social

reality. (Mannheim 1925/1952, 144)

It is slightly misleading to take the term ‘‘unmasking’’ from Mannheim;

for the word is that of his translator. The German original is enthüllung,

which means revealing or exposing. In Wagner’s Parsifal the cry goes

up, ‘‘Uncover the grail!’’—Enhüllet den Gral! ‘‘Unmasking’’ has, in ad-

dition, an overtone of exposing something that was deliberately covered,

in order to conceal its true nature. The ‘‘unmasking turn of mind,’’wrote

Mannheim, is

a turn of mind which does not seek to refute, negate, or call in doubt

certain ideas, but rather to disintegrate them, and that in such a way

that the whole world outlook of a social stratum becomes disintegrated

at the same time. We must pay attention, at this point, to the phenom-

enological distinction between ‘‘denying the truth’’ of an idea, and ‘‘de-

termining the function’’ it exercises. In denying the truth of an idea, I

54 TOO MANY METAPHORS

still presuppose it as ‘‘thesis’’ and thus put myself upon the same theo-

retical (and nothing but theoretical) basis as the one on which the idea

is constituted. In casting doubt upon the ‘‘idea,’’ I still think within the

same categorical pattern as the one in which it has its being. But when

I do not even raise the question (or at least when I do not make this

question the burden of my argument) whether what the idea asserts is

true, but consider it merely in terms of the extra-theoretical function

it serves, then, and only then, do I achieve an ‘‘unmasking’’ which in

fact represents no theoretical refutation but the destruction of the prac-

tical effectiveness of these ideas. (Mannheim 1925/1952, 140)

Mannheim’s model was Marxian, and he thought in terms of unmasking

entire ideologies. He had, moreover, a sort of functionalism in mind. An

ideology would be unmasked by showing the functions and interests

that it served. Yet unmasking, in very much the terms used by Mann-

heim, has broader implications.

Mannheim wrote that the hidden history of the unmasking turn of

mind ‘‘still calls for more exact investigation’’ (141). There is a lot of

not-so-hidden history, featuring such household gods as Hegel, Marx,

and Freud. An instructive hidden history would take in not only the

unmasking of ideologies, but the local unmaskings attempted by Ber-

trand Russell and his admirers. The Russellian doctrine of logical con-

structions did not in general aim at refuting claims about theoretical or

abstract entities, but instead tried to remove extra-theoretical presup-

positions of statements about them.

Constructionism today is usually a more local sort of unmasking than

Mannheim had in mind. Undoubtedly, studies of the construction of

gender want to unmask an ideology. But let us turn to a more typical

and less discussed example.

Serial Killers

Here is a set of common beliefs about serial killers. Serial murders are

monstrous—far more crimes thus classified occur in the United States

than elsewhere—the number of serial killers has been on the rise in

many countries—serial killers are rare nonetheless—most but not all

serial killers are men—these murderers had vile childhoods—their vic-

tims are chosen at random from a specific class of hapless people (pros-

TOO MANY METAPHORS 55

titutes, black homosexuals, or whatever)—serial murder involves

warped sex.

Every one of those beliefs is widely held. Each is, in the main, true.

Together they form objective knowledge about a class of crimes, estab-

lished by experts. Or so we think. Then we come across Philip Jenkins’s

Using Murder: The Social Construction of the Serial Killer (1994). We

know what to expect. The author will not strictly refute our beliefs. But

he will teach how the classification has been made up. He will show

that the categorization of certain crimes as serial homicides functions

for the benefit of some elements of law-and-order enforcement, and he

will tell us how a new kind of expertise has come into being.

The effect of this is somewhat unsettling. It is not at all clear what to

do, or that anything should be done. Take this true anecdote: a successful

free-lance businesswoman told me that she will not let a courier with a

package into her premises, especially when her attractive young assis-

tant is about. There are too many serial killers out there. Her office is

on the fourth floor of an upmarket mixed-use building in central, well-

ordered Toronto. What is a relevant observation? At the level of truths

about serial killers: they just don’t invade premises like yours! Or at the

unmasking level: you have somehow been conned into an irrational fear

about a kind of person, a category constructed in order to serve certain

interests, and to gratify certain fantasies! The anecdote is of no moment

except as example. There may be straightforward political conclusions

to draw from unmasking. Insofar as serial killing is an especially Amer-

ican conception (the British rippers and notorious Russian and Italian

examples notwithstanding), is its ‘‘extra-theoretical force’’ intended to

deflect attention from gun control, inner-city mayhem, and so forth?

What sorts of things are, in general, to be unmasked? Above all, the

unpleasant—disaster (Fowlkes 1982). Even when we pass from specific

kinds of people, such as serial killers, to quite general attributes of peo-

ple, we are not surprised to find the construction of anger (Miller 1983),

or both danger (McCormick 1995) and dangerousness (Webster et al.

1985). The construction of joy or tenderness would astonish us. But the

all-too-good are doubtless in for it: we would be disheartened, but not

shocked, by a construction analysis of Médecins Sans Frontières. When

I first wrote the previous sentence I added ‘‘or Mother Teresa.’’ Hardly

had the ink dried than there appeared Christopher Hitchens’s (1995) sar-

donic book about the saintly lady.

56 TOO MANY METAPHORS

Hitchens did not exactly expose her—another reason that ‘‘unmask-

ing’’ is to be preferred to the original German word, which can be trans-

lated as ‘‘exposing.’’ Unmasking is different from exposing; they work

at different levels. When the American evangelist Jim Bakker was shown

to be sexually involved with acolytes and to be salting away a fortune,

he was exposed, not unmasked. The difference between unmasking Te-

resa and exposing Bakker is analogous to Mannheim’s distinction be-

tween challenging the extra-theoretical effectiveness of a doctrine, and

simply refuting it, showing it to be false.

Refuting

Mannheim distinguished refuting from unmasking. Refuting a thesis

works at the level of the thesis itself by showing it to be false. Unmask-

ing undermines a thesis, by displaying its extra-theoretical function.

The distinction is not all that sharp, for some analyses that chiefly aim

at refuting or discrediting may gain added cogency by showing how what

is to be refuted or discredited was constructed in the first place.

The construction metaphor is severely weakened by not distinguish-

ing pure cases of unmasking from mixed cases of unmasking and refut-

ing. Two remarkable books by Donald MacKenzie illustrate the differ-

ence. His Inventing Accuracy, subtitled An Historical Sociology of

Nuclear Missile Guidance (1990) unmasks, but it also refutes the claim

of any cold warrior (or of today’s sons of cold warriors) to have ‘‘cor-

rectly’’ defined missile accuracy. The measured comparisons of ‘‘our’’

with ‘‘their’’ missiles were proposed in order to satisfy various political

or technical agendas.

The point is not that missiles are not sufficiently accurate to be lethal.

The point is that exceedingly delicate, competing, and incompatible

measures of accuracy are defined to cater to two distinct interests. The

paymasters and the public must be convinced that our missiles deliver

excellent accuracy per dollar, but also that enemy missiles are so ac-

curate that we need to build yet more missiles, or else introduce mul-

tiple entry missiles that leave a large enough footprint (the jargon of the

trade) to cancel out inaccuracies. MacKenzie’s historical sociology

shows how experts and the lay public are taken in by the assertions of

the weaponeers, engineers and policy makers alike. We walk away from

MacKenzie’s book knowing that in terms of the accuracy debates them-

TOO MANY METAPHORS 57

selves, the standard measures of accuracy correspond not to some ideal

measure of accuracy but to the interests of the parties involved. The

measures are better or worse insofar as they serve the goals of maintain-

ing or expanding arsenals.

Contrast MacKenzie’s Statistics in Britain: The Social Construction

of Scientific Knowledge (1981). This is a fascinating account of how sta-

tistical knowledge was produced in order to satisfy certain class interests

of Victorian and Edwardian England. Eugenics became a dominant re-

search interest of the later part of that period, and was featured, in a

major way, in the contributions of such influential pioneers as Francis

Galton and Karl Pearson. But we do not leave this book with the sense

that regression, correlation coefficients, or the chi-squared test have

been refuted. They may still be abused. We know from Herrnstein and

Murray’s The Bell Curve (1994) that the use of these tests, to pass from

correlation to causal claims about race, is alive and well. Correlation

and chi-squared tests nevertheless remain cornerstones of statistical in-

ference, and MacKenzie did not even think of dislodging them. People

who take issue with Herrnstein and Murray do not offer new statistical

technology; they claim those authors drew incorrect inferences from the

statistics.

MacKenzie’s missile book described the social construction of missile

accuracy and refuted measures of accuracy. His statistics book described

the social construction of statistical methods, and left those methods

intact. Mannheim would not have called that unmasking. If these two

books are run together as two undifferentiated works of all-purpose con-

structionism, their distinct merits and contributions are altogether lost.

I write with some feeling here, because of my own book about mul-

tiple personality. In one chapter (Hacking 1995b, ch. 9)I explained how

a certain continuum hypothesis about dissociative behavior was set in

place. It has become dogma that a tendency to dissociate—whose ex-

treme form is multiple personality—forms a continuum. I described

how this dogma became established by questionable psychological test-

ing and abuse of statistical tools. Yes, I showed how the continuum of

dissociation was constructed before our very eyes, a micro-social con-

struction of a supposed psychological fact if ever there was one. But I

also aimed at demolishing the evidence and the techniques. I hope I

refuted the claim to fact. Because of the current enthusiasm for social

construction, I have to say, pedantically, that the chapter in question

58 TOO MANY METAPHORS

was not a piece of constructionism, even though I described the willful

construction of an unwarranted ‘‘pseudo-fact’’ by a small but very influ-

ential social group of psychiatrists and psychologists.

HUMAN AFFAIRS

In Chapter 3 I turn to construction ideas about the natural sciences.

There is a body of such work, and it has recently attracted hostile atten-

tion, but it is as nothing compared to work on human affairs. Politics,

ideology, and power matter more than metaphysics to most advocates

of construction analyses of social and cultural phenomena. Talk of con-

struction tends to undermine the authority of knowledge and categori-

zation. It challenges complacent assumptions about the inevitability of

what we have found out or our present ways of doing things—not by

refuting or proposing a better, but by ‘‘unmasking.’’ One area of focus

involves people: childhood, gender, youth homelessness, danger, deaf-

ness, disaster, illness, madness, lesbianism, literacy, authorship. An-

other is kinds of person: the woman refugee, the child viewer of televi-

sion, the psychologist’s subject. There is also behavior, such as serial

homicide or white collar crime, and feelings, such as anger. We have

vital statistics and postmodernism. We can focus on these diverse ex-

amples in different ways. For example, youth homelessness is a condi-

tion; the homeless youth, or the runaway, is a kind of person.

Should we distinguish this great variety of items from kinds of inan-

imate entities, such as the quark, or knowledge about a tripeptide? Why

are people different? We get an intimation of the answer from the mo-

tivation of much constructionism. Constructionists are greatly con-

cerned with questions of power and control. The point of unmasking is

to liberate the oppressed, to show how categories of knowledge are used

in power relationships.

It is widely taken for granted in constructionist studies that power is

not simply exercised from above. Women refugees or deaf people par-

ticipate in and assist in the power structure. One hope of unmasking is

to enable the deaf or the women refugees to take some control over their

own destiny, by coming to own the very categories that are applied to

them. I used to call kinds of people, kinds of human action, and varieties

of human behavior by the made-up designation, ‘‘human kinds.’’ It is an

important feature of human kinds that they have effects on the people

classified, but also the classified people can take matters into their own

TOO MANY METAPHORS 59

hands. I called this phenomenon ‘‘the looping effect of human kinds’’

(Hacking 1995a). I now prefer to talk about interactive kinds.

The fundamental idea is almost too simple-minded. People are self-

conscious. They are capable of self-knowledge. They are potential moral

agents for whom autonomy has been, since the days of Rousseau and

Kant, a central Western value. Quarks and tripeptides are not moral

agents and there is no looping effect for quarks. Hence constructionism

applied to the natural sciences was in the first instance metaphysical or

epistemological—about pictures of reality or of reasoning. When applied

to the moral sciences, the interest must first of all be moral. Assuredly

there are infirm boundaries. The nonhuman may increasingly be in-

vested with moral qualities—species, forests, ecosystems, Gaia. Yet the

modeling of the moral remains firmly rooted in human values and the

potential for self-awareness. Although many constructionists are moved

by deeply moral concerns, all-purpose talk of social construction has

tended to deflect attention from moral issues. This is doubtless partly

because of a nervousness, noticed in some constructionists, in admitting

the possibility of the very idea of morality. But if the point of the exercise

is moral, one should not be squeamish about saying so.

THE NATURAL SCIENCES

Karl Mannheim had an attitude to physical science very different from

that of modern constructionists. ‘‘Scientific-technical thought,’’ he

wrote, ‘‘completes just one and the same system during successive

periods . . .’’

Because it is the same system that is being built up in science in the

course of the centuries, the phenomenon of change of meaning does not

occur in this sphere, and we can picture the process of thought as direct

progress toward ultimately ‘‘correct’’ knowledge which can be formu-

lated only in one fashion. In physics, there are not several different

concepts of ‘‘force,’’ and if different concepts do appear in the history

of physics, one can classify them as mere preparatory steps before the

discovery of the correct concept prescribed by the axiomatic pattern of

the system. (Mannheim 1925/1952, 170)

This attitude is characteristic of sociology of knowledge from Durkheim

through Mannheim. It took individuals trained in the sciences to apply

60 TOO MANY METAPHORS

sociology to the sciences themselves. One such was Ludwik Fleck, a

remarkable epidemiologist and immunologist, who published over 100

medical research papers, some written in the Lvov ghetto until it was

destroyed in 1943. He was a survivor. In 1935 he published a path-break-

ing book about scientific thought-styles (Denkstile), and about the ori-

gin and development of a scientific fact (Fleck 1935/1979, Cohen and

Schnelle 1986) In retrospect he looks like the first author to have had a

thoroughly ‘‘constructionist’’ attitude to scientific facts, although bless-

edly he did not use the construction metaphor. It would not have been

very apt—or literal—for his story of the Wasserman test for syphilis.

It is part of Fleck’s thesis that scientific facts exist only within styles

of thinking, a doctrine to which I am myself sympathetic (Hacking

1992b). Fleck did not allude to Mannheim, but he did write caustically

of sociologists such as Durkheim, Lévy-Bruhl, and less well-known fig-

ures such as Gumplowicz and Jerusalem: ‘‘All these thinkers trained in

sociology and classics, however, no matter how productive their ideas,

commit a characteristic error. They exhibit an excessive respect, bor-

dering on pious reverence, for scientific facts’’ (Fleck 1979, 47). The era

of excessive respect has passed! That is one reason for the science wars

of today. Scientists feel inexorable laws of nature are not treated with

sufficient respect by the sociologists. In fact, the early sociologists did

treat laws with complete respect, and accepted the background scientific

ideology without question. Only a real scientist such as Fleck could start

questioning the mystique in which he himself had been educated.

Leaving the subject of pious reverence for later, let us try to catch a

glimpse of where Fleck was directing us. Here is one rather conservative

way to understand the thrust of his and subsequent arguments. The

standard view is of science as discovery of facts that exist ‘‘in the world.’’

The world comes structured into facts. That is not a scientific hypoth-

esis. It is a metaphysical picture.

Fleck had a different metaphysical picture. He wrote of the emergence

and development of scientific facts. He did not mean just that they

emerge in human consciousness and develop in the history of science.

He meant that the world does not come with a unique prepackaged

structure. If we want an old name for this metaphysical picture, it is

nominalism.

Constructionism about the natural sciences is also, in part, a meta-

physical position. It is directed at certain pictures of reality, truth, dis-

covery, and necessity. It joins hands very naturally with what Nelson

TOO MANY METAPHORS 61

Goodman calls irrealism: not realism, not anti-realism, but an indiffer-

ence to such questions, which in itself is a metaphysical stance. Since

neither scientists nor constructionists dare to use the word metaphysics,

it is not surprising that they talk past each other, since each is standing

on metaphysical ground in opposition to the other.

Talk of metaphysics will seem, to many, a highbrow evasion of the

issues current in the science wars. On the contrary, it is a central part

of the story, and ignorance of it brings confusion. But it is only part of

the story. Already, in 1935, Fleck was challenging pious reverence for

the sciences. After 1945 there was a backlash against science itself. Sci-

ence had been at the service of the concentration camps and gas cham-

bers; only science could have created Hiroshima and Nagasaki. There

were valiant attempts to defend the value of science as a human en-

deavor. The most notable was Jacob Bronowski’s television series, The

Ascent of Man. This was shown for uplift to millions upon millions of

English-speaking schoolchildren. It began with heartfelt concern. Here

I, Jacob Bronowski, am a man whose relatives perished in the camps. I

have made a pilgrimage there. Here I, Jacob Bronowski, am a man who

helped pioneer operations research (as the theory of efficient bombing)

for the Royal Air Force during the war. I have made a pilgrimage to

Hiroshima. But I now want to restore the Enlightenment vision of sci-

ence, as one of the greatest endeavors of the human race, which shall

save us yet, when undertaken with humility. Science can be restored to

a state of grace.

There was another reaction, what Richard Bernstein names ‘‘the rage

against reason.’’ A rage against science and scientists. A rage that con-

tinued through the nuclear arms race, the Doomsday machine, chemical

weapons, ecological disaster, the silent spring, the nuclear winter. That

rage was so powerful that it needed few allies, but in intellectual and

academic circles it latched on to the metaphysics of constructionism.

That is because metaphysics can have ideological consequences. The

sciences, for some researchers, seem to involve getting to know the es-

sence of creation, the mind of God. The metaphysics of constructionism

denies that creation had an essence, or that there is a God’s eye view. It

is a threat to such a world view. Likewise, feminist critics of the natural

sciences formed alliances with constructionists, in order to undermine

the idea that the sciences must proceed along an inevitable, preordained

patriarchal track.

Constructionism about the natural sciences is not necessarily politi-

62 TOO MANY METAPHORS

cal or critical. A constructionist could be committed to the current en-

terprises of the natural sciences, and just as full of admiration for past

genius and present achievements as the most gung-ho science journalist

who weekly announces the latest discovery. But constructionism can

be used to unmask an ideology of science, an ideology that is intended

to produce pious reverence. It must be said, as a purely anecdotal gen-

eralization, that every single constructionist about the natural sciences

whom I know well is thoroughly irreverent.

The science wars, as I see them, combine irreverent metaphysics and

the rage against reason, on one side, and scientific metaphysics, and an

Enlightenment faith in reason, on the other. Hence the next chapter is

about metaphysics and rage.

Tuesday, April 30, 2024

measure: too many metaphors

Email List

Queen Kong and You're Only Massive- Dot Dash

Label Cloud