Michael Nicholson “The Scientific Analysis of Social Behavior”

Nicholson, Michael 1983. The Scientific Analysis of Social Behavior: A Defence of Empiricism in Social Science. London: Pinter.

 

The Nature of Scientific Statements

Thus our theories are not in the mind alone but also in what we presume to be an external reality, which is perceived by our senses. […] Clearly from one point of view this makes the position of science precarious in that all our theories are dependent on the conceptual framework. (37)

In talking of accepted body of theory, I shall not, unless specifically drawing attention to it, be merely making a comment about what most people working in a particular area happen to believe. I shall mean a body of theory which is highly confirmed by observation and is regarded as true, subject to the validity of the conceptual framework in which the concepts are interpreted. (38)

However, language is not a totally circular system but has contacts with the real world and some things have to be defined outside the language itself. Thus, some concepts must be defined ostensively, that is, by directly connecting the words of the language with objects or events. (38-39)

The propositions of the theory, then, consist of statements whose terms (in theories expressed in natural language, ‘words’) are understood by the readers and writers, having been defined ostensively or nominally from earlier ostensive definitions (ultimately ostensive definitions, that is they need not necessarily be ostensive within the specific scientific enquiry). (39)

A theory is a set of laws, which are general statements that purport to be about the world. To see whether a law does in fact conform with the world we cannot of course observe the law, which is a meaningless phrase, but only instances of it. (40)

A statement which is about the world asserts not only that something is true but also asserts (by implication) that other things are false. (41)

Popper complains of being misunderstood on this point but I can find no clear statement of how it is that a failed test is not a refutation and how a test can be a test if it is not possible to fail it. If these two propositions are correct, then I reiterate that testing and refutation are formally the same. (42)

No form of direct observation whatever is possible in the case of the centre of gravity, not because of any limitation in the observational tools but because of the nature of the concept itself. It is inherently not observable. Its ‘existence’ is determined only by the role it plays in a set of deductions which have observable consequences, but its meaning comes only from its position in this deductive framework. (45)

The concept is ‘invented’ rather than ‘discovered’. Something is discovered if a set of observations indicates that it exists. Its existence might well have been predicted by a theory but it could (in principle) have been discovered without the theory. However, the theoretical term is invented as a means of organizing observations, that is, as a part of a theory without which the concept is meaningless and hence ‘undiscoverable’. (46)

 

Goals and Theoretical Terms in the Explanation of Social Behavior

The goal of an action is the end which the actor intends it to produce. At some point there must have been some act of choice (conscious or unconscious) so that some alternative goal would have been possible. We can use an alternative vocabulary and talk about action being motivated. A motive is a desire to achieve a goal and on these definitions anything which is said about motivation implies some parallel statement about goals. I shall confine myself to using ‘goal’. Goals can be conscious or unconscious; from the point of view of the analysis here it is not very important. (65)

It is the goal-seeking activity which is the theoretical term. […] Thus the goal-seeking formulation of the problem can be seen as a theoretical form. By this I mean that any statement couched in this form is necessarily a theoretical statement whose meaning and definitions lie solely in the deductions which can be made from a system in which it plays part. […] The theory is derived from a set of postulates which underlie the goal of maximization of utility. Now utility, like the centre of gravity, is definable only in terms of the theory and the maximization of utility is a useful concept only because it permits deductions to be made which themselves do not contain the term. (70)

Propositions asserting goal-directed behavior are theoretical propositions. They do not themselves have empirical reference and are testable only inasmuch as they form a part of a theory which has testable lower-level propositions. Thus the concern felt over their possibly dubious position in a supposedly empirical science is for the most part misplaced. They are propositions which are of a different form from normal causal propositions, but they fit perfectly happily within a deductive system, yielding testable propositions in normal causal form. (71)

Thus, it is the decision to try to achieve the goal, and to select a particular procedure for doing so which causes the acts and not the goal itself. (73)

Three different senses of ‘unrealism’:

1) idealization. The criterion of a theory’s adequacy requires that the discrepancy between a deduction made within the theory and the corresponding statement about the world is sufficiently small for whatever purpose is in hand. The permissible degree of discrepancy is something which is given external to the theory itself and depends on its application to some problem. (75)

2) limited domain. If a model is to be useful then its domain has to be specified. The as if  procedure is no good if it is not clear what problems the assumption applies to. […] How much it matters whether an assumption is unrealistic or not depends on how wide or interesting the domain is and whether it is clear what elements are unambiguously within it. (76)

3) verbal unrealism. It is a form of unrealism which raises problems in the social sciences only in goal directed models, though on analysis the problems can be seen to be entirely trivial. An actor in a situation frequently does not use the same concepts in formulating his conduct as the analyst who describes his behavior. This is, of course, not surprising when the actor is not using goal-seeking terms. It might appear to be unrealistic when the actor’s descriptions of the situation are phrased in goal-seeking terms, but different terms to those used by the analyst. There is then an apparent conflict between the analyst and actor. (77)

There are two completely different forms of statements covered by an as if proposition. The first are the theoretical propositions which it is meaningless to test directly. The second form of as if statements are those which are actually incorrect, such as the mathematical billiards player discussed above. Within the relevant domain it might be perfectly predictive but it is false, can be directly tested, and further outside the domain will yield incorrect deductions. (78)

An organization has a procedure for taking decisions which may be formally or informally devised. Whichever it is, the organization has a decision-making structure, which the crowd has not. This means that I can avoid the problem of analyzing the concepts involved in such phrases as the ‘crowd has decided’. (80)

The issue is to what extent the behavior of the organization as a unit within its own higher-level social system can be understood without reference to the events within the subsystem. […] Another issue which is raised in systems where groups are the unit, is whether the group chosen is the ‘proper’ unit of analysis. (81)

A social unit is a model, an abstraction with a limited domain, and can be used for answering questions only within its domain. (82)

 

Artificial Worlds

It is an abstract world which exists purely in linguistic or mathematical form but where the statements involved in the model are simpler, either in their complexity or number, but similar, to statements which are presumed to be true of the real world. (108)

A theory is a set of propositions which, if it is a correct theory, are propositions about the real world and hence are facts. They can properly be tested against the real world. There are no clear criteria for the construction of such models and a good model is just one that ‘looks good’. (112)

A simulation is a physical system (this includes human systems) which is investigated not for its own sake but for some presumed parallelism between it and a primary system one wishes to investigate. (117)