1 

Natural Kinds, Human Kinds, and Historicity (1997)

Kurt Danziger

Abstract: Common psychological categories, such as personality, motivation, attitude, emotion,

do not correspond to inherent divisions within a timeless human nature. They do not represent
natural kinds, such as gold, which exist independently of how we depict them. The categories in
terms of which humans understand their individual conduct and experience are part of human
social life and change as that life changes. They are "human kinds" (Hacking) also in the sense
that humans are affected by the terms in which they understand themselves. The culturally
embedded and historically changing meaning of specific psychological categories forms a layer
of implicit knowledge usually taken for granted by more explicit psychological theorizing.

I want to begin by making a distinction between two kinds of theorizing that go on in psychology -
and any other science, for that matter. Mostly, when we talk of theorizing, we are referring to an
activity that involves explicitly formulated propositions, explicitly articulated assumptions, and
often clearly described models. However, there is another kind of theorizing that goes on out of
view and usually remains behind the scenes. It is this second kind of theorizing that I want to talk
about. In particular, I want to focus on certain presuppositions that are built into the network of
categories that psychologists use to define the subject matter of their scientific and professional
practice.

One cannot formulate psychologically relevant theories without the use of psychological categories.
Nor can one communicate one's empirical observations without falling back on a network of pre-
existing psychological categories which define what it is that is being observed. To be
psychologically interesting both theories and observations have to be couched in terms of
psychological categories. Learning, motivation, sensation, intelligence, personality, attitude,
constitute examples of such categories.

Psychologists have devoted a great deal of care to making their theoretical concepts clear and
explicit. But much of this effort has been undermined by their complaisance about the way in which
psychological phenomena are categorized. The meaning of these categories carries an enormous
load of unexamined and unquestioned assumptions and preconceptions. By the time explicit
psychological theories are formulated, most of the theoretical work has already happened - it is
embedded in the categories used to describe and classify psychological phenomena.

A century of specialized usage has not sufficed to eliminate the dependence of basic psychological
categories on shared understandings in the general culture. Psychology may have developed certain
theories about drive, about intelligence, about attitudes, and so on, but the network of categories
that assigns a distinct reality to drive, to intelligence, to attitudes etc. has been adopted from the
broader language community to which psychologists belong.

One consequence of this is a disjunction between the way scientific psychological discourse handles
explicit theoretical concepts and taken for granted psychological categories. Conventionalism
characterizes the deployment of explicit theoretical concepts. It is generally accepted that such
 2 
 

concepts are human inventions subject to continuous revision in the light of new research. However,
when it comes to the domains that their theories are meant to explain psychologists are inclined to
adopt a stance of unreflecting naturalism. They tend to proceed as though everyday psychological
categories, like intelligence, emotion or learning, represented natural kinds, as though the
distinctions expressed in such categories accurately reflected the natural divisions among
psychological phenomena. Psychological discussions typically assume that there really is a distinct
kind of entity out there that corresponds exactly to what we refer to as an attitude say, and it is
naturally different in kind from other sorts of entities out there for which we have different category
names, like motives or emotions.

The belief that scientific psychology adds to our knowledge of attitudes, drives, intelligence, etc.,
involves the implicit assumption that there is a fixed human nature whose natural divisions are
reflected in this received network of categories. A sensation is not an emotion and an attitude is not
a memory, though relationships between them are conceivable. While psychological theory
addresses at length such topics as the structure of intelligence or the laws of motivation, it quietly
assumes that the terms "intelligence" and "motivation" refer to distinct kinds that require
explanation by means of separate sets of theoretical constructs. What is certain, however, is that
psychological theory requires some pre-understanding of that which it is a theory of.

That pre-understanding has generally involved the unspoken conviction that psychological
categories constitute historically invariant phenomena of nature, rather than historically determined
social constructions. Therefore, the most appropriate way of investigating them would be by means
of the experimental method of natural science rather than by means of historical analysis.

The traditional historiography of psychology reflected these commitments. It did not question the
currently entrenched divisions among psychological domains, assuming that those divisions truly
reflected the actual structure of a timeless human nature. Though categories like "intelligence",
"personality" and "learning", may only have become psychological categories at the end of the 19th
century earlier texts were reinterpreted as though they contained psychological theories about such
topics. The timelessly true shape of such categories was assumed to be defined by present day usage
(Danziger, 1990). Older work was appreciated only insofar as it "anticipated" what we now know to
be true.

The older historiography considered only two kinds of factors in the development of a science, the
discovery of empirical phenomena and the construction of explicit theories that would account for
them. It tended to overlook the existence and historical change of categories that incorporated basic
assumptions and provided the framework which gave a particular structure to both theories and
phenomena.

One historian of science whose work ran counter to the prevailing trend was the French historian of
biology, Georges Canguilhem. Among the topics whose history Canguilhem (1955, 1979, 1989)
investigated was that of the reflex, biological regulation and normality. These are clearly not
theories, as that term is ordinarily used. One can have theories about reflexes, about biological
regulation, about normality, but these notions themselves are not theories. Nor are they phenomena.
They are categories that provide a framework for identifying phenomena, giving them a particular
 3 
 

meaning. Such frameworks are historical constructions, and it is the job of the historian of science to
trace their development.

The topics whose historicity Canguilhem investigated were biological categories. In due course,
some of these biological categories provided the basis for current psychological categories.
Examples of such categories are those of stimulation, intelligence, behaviour, and learning. These
provide a framework for describing and identifying psychological phenomenona in a certain way.
The possibility of describing phenomena in terms of such a framework did not always exist because
these categories only became part of the history of psychology relatively recently.

In a recently published book (Danziger, 1997), I traced historical changes in the meaning and use of
such biologically derived categories, as well as several other common categories of psychological
discourse, including personality, motivation, emotion and attitude. In each case I explored the
historical context in which modern psychological categories emerged and the way in which they
gradually acquired their current meaning.

When one conducts such an analysis it soon becomes apparent that psychological categories were
always relevant to the lives of those who used them, whether they were ordinary people or experts.
Changes in these lives were accompanied by changes in psychological categories. Although, in the
light of their historicity, it is difficult to say that these categories represent natural kinds, what one
can say is that they represent relevant kinds. They are relevant to the people who use them, relevant
to their concerns, their interrelationships with each other, their possibilities of action. There are
factors in their lives which lead them to make and to emphasize certain distinctions and to ignore
any number of others. That is reflected in historical changes in psychological categories.

Nevertheless, psychologists have always tended to think of the categories they employed as "natural
kinds", groups of naturally occurring phenomena that inherently resemble each other and differ
crucially from other phenomena. Psychological categories were assumed to represent natural
divisions among objective features of the world that existed independently of the efforts of
psychologists.

However, there are good reasons for rejecting natural kinds as an appropriate conceptual basis for
psychology. Natural objects, as defined by natural kinds, are indifferent to the descriptions applied
to them. If we change our identification of a chemical compound as a result of advances in
techniques of analysis, this changes our knowledge of the compound but the compound itself
remains the same compound it always was. But psychological objects behave in a very different
manner. A person who learns not to think of his or her actions as greedy or avaricious but as
motivated by a need for achievement or self-realization has changed as a person. Students who learn
to classify things they see under a microscope no longer have the same perceptual experience they
had during their initial encounter with microscopic preparations. The sorts of things that psychology
takes as its objects, people's actions, experiences and dispositions, are not independent of their
categorization.

This is hardly surprising because the individuals who are the carriers of psychological objects are
able to represent these objects to themselves in a self-referential fashion. Radical behaviourists
believe that such representations have a purely epiphenomenal status, but more generally it is
 4 
 

believed that their existence introduces a profound distinction between psychological objects and
natural objects that have no capacity for self-reference. The manner of their articulation in language
becomes a constituent part of psychological objects so that their identity changes with changes in
psychological language (Taylor, 1985).

Another reason why psychological objects are not independent of their categorization is that they
are intelligible only by virtue of their display within a discursive context. Whatever forms they
assume are due to their embeddedness in particular discursive practices (Semin & Gergen, 1990).
The conception of psychological entities as natural objects is often grounded in a naive belief in the
existence of a private world of psychological essences. However, distinctions that constitute
emotions as emotions, motives as motives, cognitions as cognitions, and so on, do not exist in some
sealed private box before they are so labelled in public. Identifying experiences, actions and
dispositions is not like sticking labels on fully formed specimens in a museum. Psychological
objects assume their identity in the course of discursive interaction among individuals.

Distinguishing among kinds of actions and kinds of people is part of human interaction everywhere.
Psychology attempts to offer causal explanations of the domains created by these distinctions, using
empirical investigation and theoretical hypotheses. The extent to which these attempts act back on
the distinctions themselves depends on the authority commanded by psychological expertise in a
particular culture. It also depends on the way in which psychological work relates to existing needs
and interests. If the work is truly innovative and threatens established preconceptions and
relationships it will meet a great deal of resistance. But the great bulk of psychological work has
never been in any danger of this fate. Both in its inspiration and in its effects it has been profoundly
conservative. Except on a very superficial level, it has shared the prevailing preconceptions of its
culture and arranged its investigations in such a way that no knowledge with revolutionary
implications could possibly emerge from them. In assessing the effect of psychological science on
psychological kinds it is easy to overlook the biggest effect of all, namely, the reinforcement of
existing culturally embedded preconceptions and distinctions.

This cultural embeddedness accounts for the taken for granted quality that so many psychological
categories possess. It is a quality that makes them appear "natural" to the members of a particular
speech community sharing a certain tradition of language usage. However, this sense of "natural" is
not to be confused with the concept of natural kinds that has featured in the present discussion.
Natural kinds have nothing to do with culture, whereas the natural appearing kinds of psychology
have everything to do with it. We need a term for the latter that will recognize this distinction. The
term "human kind", introduced by Ian Hacking (1992), is useful here. Hacking's main interest is in
categories that define kinds of people, like homosexual or multiple personality disorder, but, in
principle, kinds of human activity are covered too. The difference between natural and human kinds
rests on the distinctions I have already mentioned, that is, whether the kind is self-referring and
whether it is intrinsically part of social practice.

One consequence of the distinguishing features of human kinds is that their relationship to the
reality they refer to is different from that of natural kinds. The latter refer to something that would
be the case, whether any particular act of reference had occurred or not. Human kinds, on the other
hand, affect that which they refer to. Historically, "the category and the people in it emerged hand in
hand" (Hacking, 1986: 229). The way humans categorize themselves and their activities is not
 5 
 

independent of their actual conduct, because, as we have noted, such categorization is part of human
conduct and therefore not a matter of indifference to the people concerned. This leads to what
Hacking (1994) has described as "looping effects", the reaction of people to the classes to which
they and their activities are consigned. This reaction may range all the way from passive acceptance
to militant refusal. In other words, the meaning of human kinds develops and changes in the course
of interactions among those affected. (This interaction has sometimes been described as a process of
"negotiation", though that implies a more deliberate and more articulate process than is often the
case). Human kinds of the sort I have analysed (Danziger, 1997) are not natural kinds, but neither
are they mere legends. They do refer to features that are real. But it is a reality in which they are
themselves heavily implicated, a reality of which they are a part.

The reality to which human kinds refer is a cultural reality, and that in several senses: First, because
the phenomena depicted are ones which exist only in some cultural context; second, because these
phenomena commonly depend on a certain social technology for their visibility and their
production; third, - and this is the aspect that has been the focus here - because the categories used
in their representation are culturally grounded. Psychology has acquired its categories from the
culture that gave rise to it and in which it remains embedded.

Consequently, all psychological categories have been historically variable constructions. To gain an
understanding of the categories in common use at the moment, we need to see them in historical
perspective. When we go back to the origin of these categories we usually find that what later
became hidden and taken for granted is still out in the open and questionable. We also discover
some of the reasons why a new category was introduced and by whom. Because psychological
categories are heavy with historically formed pre-understanding one hopes that a better
understanding of their historicity will promote their more insightful deployment in everyday
practice.

NOTE

Paper presented at the International Society for Theoretical Psychology meeting in Berlin, April
1997. Originally appeared in W. Maiers et al (Eds.), Challenges to Theoretical Psychology.
Toronto: Captus Press, 1999, pp. 24-32.

REFERENCES

Canguilhem, G. (1955) La formation du concept de réflexe aux 17e et 18e siècles. Paris: Presses
Universitaire de France.

Canguilhem, G. (1979) Wissenschaftsgeschichte und Epistemologie. Frankfurt: Suhrkamp.

Canguilhem, G. (1988) Ideology and Rationality in the History of the Life Sciences. Cambridge,
MA: MIT Press.

Danziger, K. (1990) `Generative metaphor and the history of psychological discourse', in D.E.
Leary (ed.), Metaphors in the History of Psychology. New York: Cambridge University Press.
 6 
 

Danziger, K. (1997) Naming the Mind: How Psychology Found its Language. London: Sage.

Hacking, I. (1986) `Making up People', in T.C. Heller, M. Sosna, and D.E. Wellerby (eds.)
Reconstructing Individualism: Autonomy, Individuality, and the Self in Western Thought. Stanford,
CA: Stanford University Press. pp. 222-236.

Hacking, I. (1992) `World-making by kind-making: Child abuse for example', in M. Douglas and D.
Hull (eds.) How Classification Works: Nelson Goodman among the Social Sciences. Edinburgh:
Edinburgh University Press. pp. 180-238.

Hacking, I. (1994) `The looping effects of human kinds', in D. Sperber, D. Premack, and A.J.
Premack (eds.) Causal Cognition: A Multi-Disciplinary Approach. Oxford: Clarendon Press. pp.
351-383.

Semin, G.R. & Gergen, K.J. (eds.) (1990) Everyday Understanding: Social and Scientific
Implications. London: Sage.

Taylor, C. (1985) Human Agency and Language. Cambridge: Cambridge University Press.