On Science and the Study of Politics, Definition

In effect, as children of whatever language community, we all inherited this elementary, ordinary, and essentially spontaneous conceptualizing faculty. There are spontaneous and relatively unseif-conscious ordinary conceptual schemata to which we adapt— which are the patrimony left us by our antecedents—which they spontaneously evolved in order to adapt to the relatively circumscribed demands made in a particular ecological and cultural niche under general historic conditions. We (just as they did) employ language to guide ourselves through everyday activities, and in that essentially untroubled context we (just as they did) talk of external things known to us through their action on our bodies. The primitive terms of such an elementary language refer to essentially unproblematic sensory experience. As we mature and expand the range of our experience, we find ourselves outside the untroubled range of elementary ordinary language. We find ourselves in situations which make different demands upon us—in situations in which elementary common schemata are no longer effective guides. There is insistent reason for undertaking more deliberate inquiry into our untroubled and unseif-conscious talk. Such inquiry is none the worse for being undertaken initially within the spontaneous and entrenched framework of the elementary funded language. No inquiry is possible unless some conceptual scheme is initially available.4

The spontaneous, more complex conceptual language of adults in our society, on the other hand, is couched in terms of macrophysical “thing” objects enjoying multiple predicative and relational properties. We begin with a complicated but unsophisticated language structured by unseif-conscious, ordinary conceptual categories. Our use of concepts is undeliberative. Jean Piaget indicates, for example, that seven and eight year olds adequately employ complex conceptual words like “because.” Yet their use is unreflective—outside the range of familar employments (probably the consequence of simple mimicry), the use of the concept becomes bizarre. What knowledge the child has of the use of such concepts is, at best, procedural. He knows, within a familiar range of uses, how to use the concept in commonplace contexts. And such knowledge is, by and large, adequate. He has little, if any, occasion to invoke the term except in familiar circumstances. The conscious and reflective use of the term is the result of using it in new and more complex problem situations. We begin to specify its “grammar” in the measure that the linguistic operation is no longer undertaken in familiar and unproblematic contexts. To become conscious of the grammar of a term, to exhibit its function, is to transfer its use from the ordinary and familiar to the extraordinary and unfamiliar. “Because” is a complex relational concept; its use varies with its context. To understand its implications is to become reflective and deliberate in its use, to be conscious of its varying and various functions.

Spontaneous and unreflective common conceptual schemata are predictably informal. We have procedural knowledge of the use of their constituent concepts. In operating with such concepts attention is focused on the business at hand, the communication or the organizing of experience in fruitful fashion, never on the act of concept use itself. Such concepts, and the schemata in which they unseif-consciously appear, are simply “given”

Actually the term “because” invokes a complex concept, a concept similarly invoked by its analogues: “therefore,” “since,” “if. then,” “gives rise to,” and “produces,” among others. In ordinary language, even as adults, we use such signs with the ease of familiarity. We exchange assertions containing such signs and a variety of analogues as though they were all equivalent and equally transparent. It is only when assertions are challenged, and truth warrants are requested, that we begin to realize that the uses of “a gives rise to b,” “b because a,” and “if a then b” are not at all luminous. We begin to appreciate that the relationship between a and b cannot be adequately characterized in terms of “gives rise to,” “produces,” or “therefore.” Such employments conceal a host of difficulties.

The deliberative use of a complex concept like “because” begins with an effort to characterize its employment. A “because” response is a reply to a “why” question. A “why” question is the result of having observed some variation in human or natural events that occupies our attention. A reasonably discrete class of variable behavior, in the partially formalized sciences, is identified in terms of a “dependent variable.” Some variable behavior is identified as “dependent,” the consequence of the impact of some one or several operant factors. One asks why individual, collective or natural behavior varies as it does. The response to such questions is made in terms of an “independent” variable or variables effecting changes in a “dependent” variable or variables—ideally a testable hypothesis is framed to account for the perceived variation. The variable behavior of the individual voter, for example, is conceived as influenced by an indeterminate, but finite, number of factors— his intelligence, political socialization and the sum of funded information and motivation among others. One conjures up a schema in which there is a multiplicity of operative influences. One further attempts to isolate specific influences and one speaks of “intervening” variables, and “operative” variables.

Intrinsic to such efforts is a systematic and relatively rigorous effort to specify what is to count as a variable. Concepts are defined in terms of observable recognitors and, if possible, quantified to insure precision and reliability. “Growth rates,” and “frequency rates” are assigned to variable behaviors. Crime statistics, voting behavior and the prevalence and salience of attitudes are entertained in order to permit knowledge claims to be formulated with confirmable precision. The relationships between identifiable variables, dependent and independent, are characterized as reversible or irreversible, necessary or substitutable, sufficient or contingent, deterministic or stochastic, sequential or coextensive, or conceived in a complex interdependency relationship.5

When the use of constituent concepts, understood to stand in some kind of correlative connection to each other as dependent and independent variables, becomes deliberative, a preliminary or quasi-systematic conceptual schema becomes partially formalized. A concept is given (among others) lexical, criterial or contextual definition. Its use is self-consciously specified. Any systematic knowledge enterprise soon finds itself preoccupied with partial standardization of semantic meaning and partial formalization of the putative relationships that obtain between variables. The unseif-conscious ordinary language schema inherited from common discourse becomes gradually transformed, in disciplines like history and political science the process has been less evident than it has been in the maximally formalized sciences, but it is apparent nonetheless. The process is manifest in the work of individual authors.

In his youthful works, for example, Karl Marx regularly characterized the relationship between variables in terms of metaphors. “Political ideas,” for instance, were spoken of as the “efflux” or the “sublimates” of “material life activity.” The “economic base” was spoken of as “determining,” and sometimes as “conditioning,” or “altering” “social consciousness.” By the time Das Kapital was written, twenty years later, the assertions concerning “effluxes,” “sublimating,” “giving rise to,” “determining,” and “conditioning,” had become more sophisticated “tendency statements,” qualified by contingencies, and countertrends. An effort had been made, by Marx himself, to reduce the exploitation of metaphor and advance testable propositions specifying the relationships conceived to obtain between reasonably well defined variables.

The development of more sophisticated, deliberative conceptual schemata, characterizes the progressive evolution of significant inquiry. Concepts begin to be self-consciously employed. The initial move in the direction of deliberative use generally takes the form of more adequate definition. Scholars and research personnel lament the “semantic confusions,” the “obscurity, vagueness and ambiguity” which markedly interfere “with fruitful research,” competent storage, retrieval, and interpretation of information.6 The claims made via spontaneous and unreflective quasi-scientific conceptual schemata are fugitive—insulated from counterinstances and disconfirming evidence by virtue of their vagueness and ambiguity. the undeliberative schemata of ordinary language. Commonsense admonitions like “look before you leap,” and “he who hesitates is lost,” are both advanced and defended as true largely because they are couched in the vague and ambiguous language of ordinary discourse which insulates them from both confirmation and disconfirmation. Similarly, Marxist claims like “material life conditions determine consciousness,” and “material life conditions alter “consciousness” are both tendered and conceived true in the same unreflective manner. Maintaining that “material life condilions” determine “consciousness,” however, means something quite different, and requires a different truth warrant, than saying that such conditions alter “consciousness.” Nor can we say with any confidence that we understand what the complex sign “material life conditions” unambiguously refers to. Marxist enthusiasts will similarly employ the term “consciousness” as though its signification were self-evident. Ordinary language and rudimentary scientific language harbor such obscure knowledge claims because both are, in variable degree, nondeliberative and imprecise.

It is the search for reliability in assigning truth status to such claims that compels serious efforts at a more systematic exhibition of the logic of concept formation and the “grammar” of their respective uses. Generally these efforts involve a more careful characterization of the semantics and syntactics of concept employment. Signs which refer to properties or relations are accorded explicit, recursive, implicit, criterial, contextual or stipulative definition.

Definition

A definition is a way of teaching someone how to employ the definiendum, the sign to be defined. Logicians and mathematicians
speak of explicit definitions, an enumeration of those attributes conceived as necessary and sufficient to license entry into a class.
Thus the explicit definition of a “square” in formal language is: “a quadrilateral figure all of whose sides are equal in length and all
of whose angles are equal.” Such definitions are characteristic of formal, reconstructed, or artificial languages. They provide the necessary and sufficient condition for entry into a class.7 The definlens is the logical equivalent of the definiendum. Recursive definitions share essentially the same traits. They are most common in artificial languages where the definiendum is defined in terms of a finite series of transformations. They have maximum serviceability in the formalization and mathematizatiori of theory—providing specific meaning for concepts such as “optimum,” “equilibrium,” “differential,” “integral,” and the like. Outside the confines of formal or reconstructed languages, however, explicit and recursive definitions appear with remarkable infrequency.

What we generally find in partially formalized disciplines are implicit, criterial, contextual or stipulative definitions. A concept is said to be implicitly defined when the principal characteristics of the class of objects it is understood to denote are exhibited in a systematically related set of descriptive propositions which host it.

Any propositions which are statemental constituents of an empirical theory are composed of a logical and a nonlogical (or descriptive) vocabulary. The logical vocabulary can be carefully codified and the structural relations exhibited by the collection of propositions made explicit. Such a collection of systematized propositions (system defined in terms of the relationships which obtain between the concepts and statemental constituents) becomes an analytic conceptual schema. Syntactical variance is partially or maximally reduced in order to reveal the implicit semantic meaning of its constituent nonlogical concepts.

Anthony Downs’ An Economic Theory of Democracy provides instances of implicit definition. The two major hypotheses which subtend the work are 1) that political parties in a democracy share certain analogous characteristics with entrepreneurs in that they attempt to maximize the number of votes they receive in any election (entrepreneurs attempting to maximize profit) and 2) that citizens behave rationally in politics. These two hypotheses are housed in a partially formalized set of subsumptive propositions. Within the set, “party motivation” is accorded meaning in a subset of “derived” propositions about the “implications” of the primary “postulates,” or “axioms.” Downs conveniently summarized the derived propositions at the end of his treatise and provided an indication where the interstitial propositions are to be found which make the “derivation” reasonably precise. The complex concept “party motivation” is implicitly

(and incompletely) defined in the following set of propositions:
Proposition 1
Party members have as their chief motivation the desire to obtain the intrinsic rewards of holding office; therefore they formulate policies as means to holding office rather than seeking office in order to carry out preconceived policies.
Proposition 2
Both parties in a two-party system agree on any issues that a majority of citizens strongly favor.
Proposition 3
In a two-party system, party policies are (a) more vague, (b) more similar to those of other parties, and (c) less directly linked to an ideology than in a multiparty system.
Proposition 4
In a multiparty system governed by a coalition, the government takes less effective action to solve basic social problems, and its policies are less integrated and consistent, than in a two- party system.
Proposition 5
New parties arise when either (a) a change in suffrage laws sharply alters the distribution of citizens along the political scale, (b) there is a sudden change in the electorate’s social outlook because of some upheaval such as war, revolution, inflation, depression, or (c) in a two-party system, one of the parties takes a moderate stand on an issue and its radical members organize a splinter-party to force it back towards a more extreme position.
Proposition 6
Democratic governments tend to redistribute income from the rich to the poor.
Proposition 7
Democratic governments tend to favor producers more than consumers in their actions.8

A term so defined is defined via the meaning exhibited in the collection of propositions it “entails.” To appreciate the cognitive content of propositions one through seven is to know what “party motivation” means in the context of Downs’ discussion. The expression “party motivation” obviously has variable meaning in ordinary language. Downs has implicitly and partially standardized its meaning by lodging it in a logically articulated set of propositions. By doing that, Downs has provided a collection of potentially testable propositions which together provide for public meaning and set some of the principal evidence conditions governing the terms so defined.

Talcott Parsons’ treatment of “power” illustrates, similarly, the strategy of establishing and conveying meaning via the “entailments” of “implicit definition.”9 The discussion begins (as is the case with Downs) with an analogy: “Power is comparable to wealth, which, as a generalized societal resource, is allocated to many different social subsystems for ‘consumption’ or for ‘capital’ use.” From that point on, “power” is implicitly defined in the collection of propositions Parsons argues are “entailed” by the expression. “Entailment” suggests logical interrelatedness, but the logic of such definitions (particularly as they are found in political inquiry) is singularly informal and implicit definitions are, as a result, characteristically porous. Nonetheless, the strategy is that of implicit definition—essentially analytic efforts at the public characterization of meaning.

That such is the case is indicated by the fact that authors like Parsons regularly inform political scientists that their schemata “in no way” purport to be empirical contributions. Such efforts, they say, are subject to “logical, not empirical proof” —precisely what one would expect from nontheoretical, analytic formulations.

An implicit definition conveys the public meaning of a sign by housing that sign in a logically articulated set of propositions. Since the logic of such sets, particularly as they are articulated by political scientists in the areas of comparative politics, international relations and general theory construction, is not rigorous, the conceptual terms are therefore only incompletely defined.

An alternate definitional strategy regularly invoked by political scientists is that which attempts to specify public meaning via criterial or range definitions. These are definitions provided in terms of a system of overlapping and interacting criteria. No one of the criteria is in itself necessary, nor is any one of them sufficient, to provide entry into the class. The set of attributes which license entry into a class are conceived, severally, as relevant (sometimes referred to as “essential”) and, given the pragmatic intent of the enterprise in which the concept is to be employed, a judgment of “weight” can be accorded any one or any number of them. Specifying relevant attributes provides the tests which can be employed in establishing whether or not the concept is being correctly deployed over given specimens. In order, for example, to determine if a bit of behavior is to count as a “crime,” or as evidence of a “prefascist” disposition, some reference must be made to some symptomatic observable evidence that the specimen is required to display if entry into the requisite class is to be warranted.

By way of illustration one can consider the criterial definition offered by Carl Friedrich and Z. K. Brezezinski characterizing “totalitarian dictatorships.” The authors offered, synoptically, four criteria governing entry into the class: 1) an official ideology; 2) a single mass party led typically by one man; 3) a system of terroristic police control; and 4) a technologically conditioned near- complete monopoly control of the media of information, the means of coercion and the national economy. Any system of government that evinces all such characteristics is spoken of, properly, as a totalitarian dictatorship. The collection of symptomatic traits constitutes a criterial definition of a special class of objects.

Such a criterial definition can be conceived as providing a dual set of mutually exclusive and exhaustive categories: i.e, totalitarian and nontotalitarian. Such a binary system is obviously inadequate to represent political reality, and authors like Dante Germino have argued that although Fascist Italy “fell short of the totalitarian mark” with respect to the extent and degree of police control employed, the regime, nonetheless, qualified as “totalitarian.”

The strategy tacitly employed is to conceive the concept “totalitarian” in terms of criteria!, rather than explicit, definition. As defined by Friedrich and Brzezinski, “totalitarian” is an “extreme” type with a graded series of intermediate types occupying attribute space between the two extremes, “nontotalitarian,” and “totalitarian.” A precise division of the two, Germino implicitly maintains, is “artificial,” theoretically sterile, “untrue” to the political reality. The property traits of totalitarianism can, by implication, be exhibited in varying degrees, degrees measured quantitatively (e.g., what determinate degree of police control would qualify as “terroristic,” what specific measure of formal support must an ideology be given to count as “official”?) or qualitatively.

Similarly, should there be no “unitary party,” but a “popular front,” while the remaining symptomatic traits were in evidence, would the political system still count as “totalitarian” or not? That is, if any one of the criterial traits were totally absent would the system still be characterized as a member of the class “totalitarian”?

In effect, such criterial definitions in political science are more frequently than not merely programmatic suggestions, a heuristic propaedeutic to systematic inquiry. Such formulations generally constitute suggestive programmatic characterizations— mnemonic summaries of intuitive assessments of a wide variety of descriptive materials. Once the programmatic characterization is given specific reference, what results is usually a typology, an ordering of concepts of a purely comparative sort representing a continuum of complex attribute spaces.
Because of such considerations, and the frequency with which they are not appreciated, political scientists will often differ in their characterizations of systems. Germino counts Fascist Italy as “totalitarian”—Hannah Arendt, H. Stuart Hughes, and Sigmund Neumann refuse to do so. Many political scientists count the Soviet Union as “totalitarian”—Allen Kassof has suggested an alternate classification, “administered society,” to accommodate the Soviet Union as a “totalitarian” system without police terror. ‘

Criterial definitions can be effectively employed only if the criteria marking entry into the class is specified with all the precision the subject matter allows. What is, for example, “near monopoly” or “bureaucratic” control of the economy, and how is it measured? If the “unitary party” is spoken of as “hierarchically, oligarchically organized,” how is the degree of such organization identified? Each of the criterial attributes has variable manifestations. Moreover, one must decide if a “significant” or “essential” similarity obtains even if one or more criterial attributes are effectively absent. Criterial definitions, to be employed, require judicious “weighing” and specific characterization. The ultimate justification of their use will turn, minimally, on at least the following considerations: 1) are the criteria delivered with sufficient precision to render their employment unequivocal? 2) are the attributes “weighted,” that is, are some of the attributes conceived more essential to the generation of more economical and more fruitful schemata in terms of didactic, heuristic, mnemonic, or theoretical yield? The use of criterial definitions is warranted by their power, their ability to more effectively serve as instructional or recall conveniences and/or their capacity to generate more testable propositions involving a wider range of phenomena than any alternative characterization. Ideally, the most suitable characterization would achieve maximal yield in all these respects.

The talk about “essential attributes” of the phenomena under political investigation and “real definitions” which have significant theoretical import, generally alludes to criterial definitions of the sort indicated above.12 Such definitions are proffered as a consequence of a search for an empirical explanation of some phenomenon or set of phenomena. The inquiry generally begins with concepts borrowed from ordinary language. Their employments in ordinary usage are, however, so vague and ambiguous that their serviceability in any rigorous inquiry is, for all intents and purposes, precluded. In the effort to render their meanings more precise, the investigator undertakes an explication, an enterprise involving judicious synthesis and rational reassessment, employing criterial or implicit, rather than explicit or recursive, definition.

Such an explication must permit us to reformulate, in propositions systematically related in relatively precise form, at least a large part of what is customarily expressed by means of the concepts under scrutiny. Moreover, such concepts permit the development of a comprehensive and rigorous theoretical system. Explication attempts to reduce the semantic limitations, vagueness and ambiguities of expressions like “near monopoly control,” “bureaucratic,” and “hierarchically organized,” for the purpose of generating hypotheses that can be syntactically integrated into a theory having explanatory and predictive character.

Such concepts, even when they are housed in a reasonably well articulated theoretical system, are open-textured. As the theory itself finds increasing application, its constituent concepts will take on extended meaning. These meanings are by-products of the theory’s application and are exhibited in the findings that the theory itself was instrumental in revealing. The porosity of criterially and implicitly explicated concepts (what Hempel and Kaplan refer to as the “openness of meaning”) refers to the fact that such concepts employed for empirical purposes have their meaning progressively elaborated in practice—and that closure is never definitive. The explication approximates, asymptotically, specific meaning. The fact is that in employing a concept in political science we are often faced with the reality that we do not know where to draw the line, for example, between “totalitarian” and “nontotalitarian” political systems (to use only one illustration). A point can be arbitrarily stipulated, but it remains true that for theoretical purposes we may find it advisable to treat the distinction as problematic, the kind of issue that cannot, in principle, be settled beforehand, once and for all.

Premature closure in the explication of meaning can, in fact, impair empirical inquiry. What generally happens is that a stable core meaning is provided for concepts having high yield potential. As inquiry progresses, meaning becomes increasingly specific. Intrinsically vague concepts, having unpredictable use variance, can have little substantive yield in inquiry. But concepts whose meanings are arbitrarily and unalterably fixed can be stultifying. Much of the research in contemporary political science is exploratory. As a consequence concepts are, at best and most frequently, formulated in some preliminary fashion which permits data to be collected, classified, processed, interpreted, stored, and communicated in responsible fashion. One attempts to characterize (i.e., “operationalize”) concrete processes and activities implied by the phenomena under observation. The process by virtue of which such “operationalization” proceeds cannot be fully characterized. It involves conscious and unconscious appeal to funded knowledge, familiarity with the specialized and general literature devoted to the domain of discourse. It involves informal logical treatment and an imaginative qualitative analysis of observational data.’4 One formulates a preliminary conceptual scheme in which the criterially or implicitly defined concepts (generally originally borrowed from ordinary speech) occupy a place. Each concept is accorded a criterial or implicit, but open-textured, definition (on the one hand, indicating a relatively stable collection of observable attributes at least one of which must obtain if the concept is to be used felicitously, or on the other, generating one or more testable attributes that would tie the concept down to the empirical world) in order to organize experience effectively. This schema, and the criterially or implicitly defined concepts which inhabit it, is then applied to the complex world of political phenomena. In finding vagueness or ambiguity in them which reveals itself in slackness of fit, redefinition of the concepts and rearticulation of the schema takes place. Since experience is ongoing, redefinition and rearticulations are potentially infinite undertakings, there are no terminal contexts of inquiry in empirical science.

This cannot be understood to suggest that the concepts employed in political inquiry should remain vague and ambiguous. All that is intended is a recognition that while all concepts in empirical inquiry must be possessed of determinate or determinable semantic meaning—that meaning remains, in a real sense, porous. Further inquiry will add unforeseen qualifications to the employment of criterially or implicitly defined concepts. This is as true for dispositional concepts, defined via observational evidence sentences (i.e., “autocratic” is a dispositional attribute when ascribed to “government” and is defined in terms of the behaviors of individuals occupying strategic or key roles in the system; but what behaviors are to be construed as symptomatic of autocratic government, or what will constitute strategic or key roles, may subtly vary in diverse contexts), as it is for descriptive concepts like “democracy” or “fascism” defined in terms of an open, but defining, set of attributes.

Contextual definitions share some of the species traits of criterial definitions. A contextual definition is one “which introduces a symbol s by providing synonyms for certain expressions containing s, but not for s itself “‘ Thus Nelson Poisby defines “power” in the following way “Power is the capacity of one actor to do something affecting another actor, which changes the probable pattern of future events.”6 The propositional form of such an elliptical contextual definition might be adequately represented as, “x stands in relationship R to y = Df, x changes the Q of y, but y does not change the Q of x.”

One possible instantiation of this construed propositional form might be rendered in something like the following contextual definition: “Power is the capacity of one actor (or group of actors) to do something affecting another actor (or group of actors), which changes the probable patterns of specified future events in circumstances in which the latter actor’s (or actors’) capacity to so affect the former, in the same sense or manner, is relatively more restricted, minimal or null.” Thus R. H. Tawney defines “power” as “the capacity of an individual, or group of individuals, to modify the conduct of other individuals or groups in the manner which he desires. . . .“ Lasswell and Kaplan define “power” as “participation in the making of decisions: G has power over H with respect to the values K, if G participates in the making of decisions affecting the K-policies of H.” “Power” can then be defined in terms of the relative frequency of success an individual or a group enjoys in attempting to have his or its purposes achieved. The frequency of success can be measured quantitatively and a “power index” provided—as has been attempted by various game theorists—in a relatively formal manner. Alternately, authors like Robert Dahi can provide signal illustrations of the interplay of interest groups to display the sharing of power or the relative predominance of power.

Contextual definitions are given in relational terms rather than in terms of properties or attributes possessed by an individual or group of individuals. “Power” itself is a nonobservable entity. It is defined contextually—in terms of relations—in a special form of dispositional analysis. “Power” manifests itself in terms of observable success in influencing outcomes. in its absolute form it is an irreversible relation between two terms. In its nonabsolute form it can be measured in a ratio of success to failure. The concept “power” is not, in and of itself, defined. What is provided are expressions which can be substituted for expressions containing the term “power”—but the concept itself is not assigned properties. Criterial definitions require appeal to at least one of an open, but finite, set of dispositional attributes manifested in observable behaviors. Contextual definitions refer to dyadic, triadic or n-adic relations—they do not appeal to possessed attributes.

Such definitions abound in political science literature. “Government” is defined as “the exercise of imperative control within a definite territory, and within that territory it successfully claims a monopoly of the use of force.” “Political elite” is defined in terms of that collection of individuals which “comprises the power holders of a body politic.”

Like criterial definitions, such contextual definitions support empirical inquiry in that they suggest testable hypotheses which warrant their application to any specific case. Since they do tend to support empirical inquiry they are, like criterial definitions, porous, open-textured, in the same sense and to the same degree. Because they support empirical inquiry they are “real” rather than nominal definitions. To know the meaning of such concepts suggests how one might determine the truth of any assertion in which they appear.

A stipulative definition, in turn, is a definition which arbitrarily restricts the meaning of a sign to that meaning assigned to it by the author. Thus Harold Lasswell and Abraham Kaplan stipulatively define “interest” as a “pattern of demands and its supporting expectations,” a meaning sufficiently distinct from ordinary usage to indicate its stipulative character. 1 Such a strategy is undertaken as a significant cognitive assist. It can either constitute a mnemonic convenience—providing a meaning suitable to ready storage and retrieval—or it might be undertaken out of reasons of economy— such usage might be conceived as reducing redundancy or duplication, or alternatively, it might provide simpler logical derivations. It might be advanced because it is understood to generate an increased number of testable propositions. Or it might be understood to possess special heuristic significance, as being capable of suggesting a number of directly or indirectly testable propositions that might not have been generated as long as the sign or sign complex remained confined to ordinary usage. In effect, a stipulative definition is introduced because of its potential theoretical yield. That is the motive which prompts such a strategy and that constitutes its ultimate justification. Such stipulations may, in fact, outrage ordinary employments and may be counterintuitive. They, nonetheless, have ample justification and political scientists use stipulations with regularity.

The strategies we have considered are those which are primarily research oriented. They are most frequently employed by practitioners concerned with producing truth warrants for complex empirical claims. The distinction frequently made between “real” and “nominal” definitions refers to the distinction between efforts at serious empirical employments and those which represent notational conveniences. Nominal (or lexical) definitions can best be, for our purposes, conceived as conventions which merely introduce an alternative notation for a given linguistic expression. A specified expression, the definiendum, is construed to be more or less synonymous with a simple or complex expression, the definiens. Nominal definitions are conveniences; they are normally abbreviatory, providing for more parsimonious speech acts. Thus H. V. Wiseman identifies “the common orientation of two or more people” as “culture,” and “that organized sector of an actor’s orientation which constitutes and defines his participation in an interactive process” as a “role.” Such definitions constitute primarily notational conveniences. They facilitate communication by rendering speech stenographic and more economical. Instead of repeating the complex phrase, “that organized sector. . . and so forth,” one simply employs the conventional term, “role.”

Such terms are, for the purposes of verbal “theory,” eliminable by substituting the definiens for the definiendum throughout. Signs introduced via nominal or lexical definition can be dispensed with, by substitution, from any context in which they appear. Everything said in the abbreviatory notation can be said, presumably, in terms of the (characteristically) longer expression which is its lexical definiens and vice versa.

As distinct from “real” definition, “nominal” definitions can be conceived as those definitions which serve, primarily, as conveniences. They are conveniences in that they attempt to make more explicit the thought of their author. They do not, in and of themselves (and this distinguishes them from stipulative definitions), suggest empirical inquiry that would license their use. They are, at best, efforts to employ terms with a relatively specific meaning in the interests of economy and consistency. Thus, when Lasswell defined “the environment of an act” as “the events other than the act itself within which the act is included,” he indicated that such a definition was “not intended to be fully precise and rigorous from a strict logical viewpoint, but only to make more explicit than is usually done the framework relating the various concepts employed.

George Homans characterizes such definitions as “nonoperating definitions”—”nonoperating” because the concepts they identify are not defined in such a manner as to generate testable propositions in social science.2 1 Such definitions are linguistic conveniences. A “role” is “the behavior expected of a man occupying a particular social position.” “Culture” is “the funded pattern of interpersonal expectations and belief systems transmitted by a society.” What behavior constitutes instantiation of “role behavior” is unspecfled. What expectations and what ideational elements are to be admitted as instances of “culture” is equally unclear. One need but compare Bertrand Russell’s lexical definition of “power” as “the production of intended effects”22 to realize its nonoperative character. Knowing that power is the production of intended effects, we know nothing other than a verbal convention. There is no way in which “power” so defined could, in itself, be made the subject of empirical inquiry. Thus we find Karl Deutsch moving from a lexical definition of power: “By the power of an individual or organization, we . . mean the extent to which they can continue successfully to act out their character,” a nonoperating definition, to a partially operationalized contextual definition: “Power. . is conceived on the analogy of the hardness scale of minerals, of the scratching of glass by a diamond or of the pecking order’ in a chicken yard,” and finally to an elaborate real definition that permits of substantial specification and quantification of related, directly or indirectly observed, variables.

The shift from “nominal” or lexical definition to “real” definition generally marks the transition from verbal “theory” to verificational strategies, that is to say, the generation and confirmation of testable theories and empirical explanation. The difference is intuitively obvious when one recognizes the nonoperative character of a definition of “power” that characterizes it as the “production of intended effects” as compared with its assessment in terms of a measurable ratio of success to failure in situations of competitive negotiation.

“Nominal” definitions can appear in the guise of verbal synonymies, criterial or contextual definitions. Characterizing them as “nominal” simply means, in the context of our discussion, that they are nonoperative, nontestable, because of the lack of specificity or behavioral reference that characterizes them.[1]

On Science and the Study of Politics

For at least a generation the putative differences that have separated the “behavioralists” and the “anti-behavioralists” in political science have been a public scandal. Robert Dahi’s urbane and sanguine anticipation concerning immanent rapprochement in the discipline1 has not been, in fact, realized, if one measures it against the abundance and the exacerbated quality of the critiques of “behavioralism” that have appeared in a variety of places. The critiques of Russell Kirk, Mulford Sibley, and the collective criticisms contained in Herbert Storing’s Essays on the Scientific Study of Politics followed close on the heels of DahI’s account.2 The relatively recent publication of critical assessments like that of Christian Bay and K. W. Kim, conveniently bound in a volume dedicated to “a critique of behavioralism,”3 indicates that the issues that divide “behavioralists” and “nonbehavioralists” are far from resolved. This material, conjoined with earlier critiques like those collected by Helmut Schoeck and James Wiggins, 4 provides a broad and loosely jointed collection of objections to the “behavioral approach” in the contemporary study of politics. These objections range from suggestions that behavioralists are, in some vague sense, crypto-Marxists and/or “characterized by conservatism” 5—that they practice a science predicated on a “dogmatic atheism” and/or that they invoke a science that lacks “even dogmatic atheism to fall back upon”6—that they are incorrigibly “holistic” or “collectivistic” and/or that they are irretrievably “methodological individualists”7 —to the insistence that they be more “unscientific” and less rigorously scientific in their inquiries and/or that they are too unscientific, and fail, as a result, to be “scientific.”

The fact that such objections enjoy evident popularity and that the charges with which they are freighted recur with such insistent regularity indicates that the strife between “behavioral” and “non- behavioral” persuasions has hardly diminished since DahI’s hopeful assessment of almost a decade ago. That many of the charges are mutually exclusive, indicates that the discussion is beset by regular equivocation and/or analytic and interpretive error.

The principal contention of this book will be, implicitly, that once certain critical concepts are at least moderately well characterized, most of the putative issues dividing parties in the exacerbated debate dissipate themselves; they are in a real sense “pseudo-problems,” the consequence of linguistic, analytic, conceptual, and procedural confusions. In effect, once terms like “science,” “truth,” “knowledge,” “understanding,” “philosophy,” “meaning,” and “values” are specified with minimal precision, few, if any, substantive issues remain to divide our discipline into dysfunctional factions.

Science and “Behavioralism”

The fact that both the critics and protagonists of”behavioralism” have rarely attempted to explicate concepts critical to the discussion of the relation of “science” to political inquiry is, at best, curious. It is curious because much of the criticism turns upon the employment of such terms. For example, “anti-behavioralists” frequently simply and flatly oppose the application of the “scientific approach” to the study of politics, or by entailment insist that its study be pursued by auxiliary “unscientific” techniques.9 As a matter of fact, what such critics frequently adduce as evidence that political

scientifically impoverished. Under such circumstances the criticism that political science requires “unscientific” supplements collapses into a confused admonition that political scientists should apply only responsible, accredited, and appropriate scientific techniques to the study of politics. In order to make such criticism coherent, what would be required is an antecedent discussion of what “science” is understood to be. Before anyone can recommend supplements to “science,” it seems reasonable that he should offer some account of what he conceives himself to be supplementing.

There is yet another reason for undertaking some preliminary effort to characterize what “science” might be. If one seeks to identify those political scientists that are understood to be “behavioralists,” it becomes immediately obvious that the term is painfully vague. At best, the term can be given a range or criteria) definition in terms of a set of overlapping and interacting criteria by virtue of which the members of the class might be identified. To be admitted to class membership, an instantial case must be in possession of at least one of the requisite criteria for admission, but the possession of no one specific trait is necessary for entrance. Any one of a number of possible traits would seem to qualify one for entrance into the class of “behavioralists.” The range or criterial definition of the term reflects the complexity and continuous variability of the membership criteria as well as the lack of specific class bounda“Behavioralists” have been identified by the possession of any

one of a number of traits. A “behavioralist” has been identified as a political scientist who 1) publishes in relatively specialized journals such as Public Opinion Quarterly, World Politics, American Behavioral Scientist, and Behavioral Science, and/or 2) enjoys membership or participation in the Social Science Research Council’s Committee on Political Behavior and the Committee on Comparative Politics; and/or 3) is preoccupied with the study of small group and voting behavior; and/or 4) undertakes research efforts conducted with a concern (a) for searching out regularities, discoverable uniformities in political behavior expressed in lawlike statements that could be, ideally, systematically interrelated to produce explanatory and predictive accounts, (b) for establishing appropriate verification techniques, (c) for rigorous scrutiny of procedures involved in the collection, processing, and interpretation of data, (d) for quantification whenever possible, relevant, and meaningful, (e) for maintaining the logical distinction between factual knowledge claims and normative prescription and proscription, (J) for a self- conscious and systematic interrelationship of theory and research, (g) for the development of explanatory and predictive theory as a necessary antecedent to applications, and (h) for the integration of the findings of political science with those of the related social sciences.

The possession of any one of these characterizing traits would admit one into the class of “behavioralists.” The “paradigm behavioralist” would be one who exemplified all these characteristics. There are, of course, few such political scientists. In general, a political scientist identified as a “behavioralist” would display only some one or some combination of these traits and the traits most generally taken to characterize the class are, in fact, those associated with a research concern directed toward discovering regularities governing political behavior, regularities that might be articulated into a systematically related collection of propositions having explanatory and predictive yield. Since just such concerns have been understood to loosely characterize “science,” it seems that the disposition to employ the scientific method has been generally assumed to characterize “behavioralism” and “behavioralists.” If such is the case—all the more reason to attempt some assessment of what “science” and its “method” might be reasonably understood to be.

Some of the critical issues that seem to divide political science as a discipline into factions turn on what science is understood to be. First, the suggestion that “unscientific” methods are essential to political inquiry requires that we have at least some understanding of what “science” might be—and secondly, the identification of “behavioralists” seems, generally, to turn on the disposition of some political scientists to attempt to employ the methods of science in their work. Science, and the methods of science, should be, therefore, the first objects of analysis if one wishes to assess the cognitive merits of the “behavioralist-non-behavioralist” dispute.

The effort to unpack the concept “science” and to offer a summary account of “the scientific method,” is beset, of course, with innumerable difficulties. The term “science” obviously has persuasive force, either positive or negative as the case might be. It has also enjoyed a long history. its specific cognitive meaning is, as a result, difficult to isolate, not only because of the persuasive employments of the term, but because the term has had a long and confused philological history. It has been deployed throughout the history of its use over such a variety of activities and such disparate collections of methods and putative truths that any significant characterization must, of necessity, be at least substantially stipulative and historically contextual.

Once this qualification is made, it can be argued that the term “science” has taken on a relatively (but not undisputed) specific contemporary meaning: it refers to those procedures, which, as a matter of historic fact, have provided a systematically articulated and comprehensive body of maximally reliable knowledge claims that afford men survival and adaptive advantage by affording explanatory and predictive leverage. 1 2 Such an explication thus includes both a consideration of “science” as process (that is to say, the term refers to methods and procedures invoked to provide for maximally reliable truth ascriptions) and “science” as product (that is to say, the term has as referent that corpus of linguistic entities to which truth can be ascribed with maximal domain variant reliability). The former is commonly spoken of as the “scientific method,” and the latter “scientific knowledge.” Both are historic products and both are historically relative.

What this implies is that neither specific observational nor experimental procedures, nor methods of generalization, nor specific logicodeductive strategies, nor any collection of procedural assumptions or presuppositions, nor any technique of measurement or instrumentation is cibsolutelv essential to the method of science. Furthermore, no single existential assertion delivered by scientists, nor any conjunction of such assertions, is absolutely essential to the corpus of science. Science is neither a specific collection of procedures nor a specific body of essential truths. Every truth warranting procedure and every truth claim remains, in principle, subject to review—none are specific to science.

Mathematicians do not make observations; geographers are but little concerned with formulating lawlike generalizations; astronomers undertake few, if any, experiments; meteorologists make only the poorest predictions and botanists rarely measure— and yet all are scientists. The products of science have, at various times, characterized our world as being composed of an infinite number of qualitatively identical, irreducible, variously shaped, ceaselessly moving, impenetrable particles, and at other times as being composed of enormously complicated systems of discontinuous positive and negative electrical charges in Hubert multidimensional space. Science has conceived our spatial environment as Euclidean and then sometimes as non-Euclidean, our local celestial environment as geocentric and then again as heliocentric. Science has argued that man himself is an intricate reflex mechanism, and again (at the same or at other times) that he is an autogenically motivated complex of physical and psychic needs.

In effect, processes and products are identified as “scientific” only by virtue of that system of overlapping and interacting criteria that evidence the complexity and the changing substance of the object of inquiry.

The one feature that we have gradually come to realize as constitutive of science in whatever form it has taken, a feature pervasive of science both as process and product, is its necessary concern with reliability. Without minimal reliability science is impossible, because without minimal reliability interpersonal communication becomes all but inconceivable. Without reliability argument and deliberation cannot proceed. Knowing anything becomes impossible.

Only when a descriptive term is deployed over particulars reliably can we understand assertions about its past, present, and future employments. Without reliability, explaining the past, organizing the present, and anticipating the future becomes impossible and rationality itself is, for all intents and purposes, abandoned. it has been the insistent awareness of these considerations, the recognition of the critical function of reliability, that has motivated the search for verification procedures of maximal reliability with respect to descriptive claims and absolute reliability with respect to formal truth claims. In the continuing effort to maximize still further the reliability assigned to descriptive claims and expand the power, range, and effectiveness of our logical apparatus, the procedures employed by science will be constantly subject to review— and with them their linguistic products and the truth status attributed to them.

Our preoccupation with reliability is dictated by the generic human concern with adapting to, and controlling, our complex and changing environment. We find it necessary to anticipate futures—and we do that best by exercising the imagal trial-and-error techniques, and their subtending rationale, that we identify as rational behavior. That behavior necessitates a sensitive awareness of verification procedures for empirical knowledge claims and an appreciation of the formal and informal logic of cognitive discourse.

Science as “Ideology”

Once science is so understood, as a dynamic and self-corrective concern for reliability in procedures and in cognitive yield, it becomes evident that the disposition on the part of some commentators to refer to science as an “ideology,” the equivalent of any alternate “ideology” (an equivalence which is a calculated criticism of science), leads not only to confusion, but to serious error as well.

An “ideology,” understood in any meaningful sense, is a system of argued beliefs having both specific empirical content and normative intent. Certainly science, at any particular time, constitutes an argued belief system, but science is intrinsically self-corrective and has no essential substantive content. Not only can science alter both its procedures and its products and still remain science—it must do so on pain of ceasing to be science. An “ideology” possessed of the same properties would simply not be an ideology. For every ideology there is some particular set of substantive propositions without which it would cease to be that ideology, if one rejected, for example, the thesis that race was a significant historical, social, and political variable in any sense, one could not entertain a National Socialist ideology. If one rejected class as a serious historic, social, and political variable, one could hardly entertain a Marxist ideology.

An ideology contains at least some identifiable descriptive propositions that are essential to its continued existence as that ideology. Without that specific content the ideology would cease to be that ideology.

Science, on the other hand, is committed to the recognition that each and every descriptive proposition, no matter how well-entrenched in contemporary science, is nonetheless, in principle, corrigible —and is therefore not essential to the continued existence of science. In this sense, science cannot be an ideology. Science is, in principle, animated by any method that can show itself to be maximally reliable, and composed of all and any linguistic products that are the consequences of the application of those methods. If intuitions, mystic insights, the dialectic, phantasy, or whispered intelligences from God proved to be maximally reliable in permitting men to empirically adapt to and effectively control their environment, they would become, ipso facto, constituent procedures of science.

The “objectivity” of science implies no more than the employment of techniques that assign maximally reliable truth status to propositions. Its propositions are “objective” insofar as they are the consequent products of the employment of those techniques. Objectivity means maximal reliability with respect to empirical truth claims and demonstrative certainty with respect to formal or analytic claims. Intersubjective observations, direct or indirect, have provided, in fact, the most reliable warrants for the first, and valid arguments the validating warrant for the second. Children and primitive peoples very quickly learn the advantages of intersubjective checks on subjective impression. Intersubjectivity, in effect, was one of the first and remains one of the necessary conditions of reliability. Intersubjective observaüon, direct or indirect, has entrenched itself as the maximally reliable method of empirical truth determination, Among all candidate alternatives it has minimized the likelihood of error. Empirical inquiry is not the kind of pursuit in which error can be made logically impossible. We can never attain absolute reliability concerning any material truth. But we can maximally reduce the possibility of empirical or factual error—and intersubjective observation is the necessary condition for accomplishing precisely that.

When propositions are articulated into arguments, the best defense we have developed against flawed conclusions is the employment of the techniques of formal logic. Every child who awaits to have his identifications confirmed or corrected, every scientist who awaits experimental replication and confirmation of his findings by suitably trained peers, has committed himself to the consistency criteria of contemporary logic, and consistency reveals itself as the necessary condition of reliability—a domain invariant requirement.

Knowledge claims can only be made in a natural or reconstructed language and language use, when employed for cognitive purpose, is rule governed. As early as antiquity the first efforts were made to standardize just those rules. Aristotelian syllogistic logic was the first major effort to codify them. It was understood that if language was to be used with cognitive intent, consistency in use was a minimal, but necessary, requirement. If prediction was to be of some effect, if generalizations were to serve orientation and adaptation, the general terms which denoted identifiable classes of particulars had to evince consistency. If generalizations held conjointly with some specific and specifiable particulars were to provide explanatory and predictive leverage, those procedures which permitted licit transit from premises to valid conclusions had to be specified. Aristotle’s syllogistic was the first systematic effort to provide a science of the implications of sentential forms. His efforts were supplemented by the Megarians and the Stoics who delivered what is now sometimes called “sentential logic,” the study of logical connectives, particles such as “and” and “or” by virtue of which sentences could be legitimately combined to form new sentences while preserving their truth status. The irregular development of logic proceeded throughout the medieval period until Gottlob Frege developed, at the end of the nineteenth century, the logic of functions or predicate logic, the study of all logical particles.

Since then the developments in logic have provided science with a rigorous technique for insuring consistent articulation and valid derivation without which contemporary physics and computer employments, for example, would be impossible. Formal logic has grown into a science of which the old traditional logic forms only a fragment. It is now a science of such enormous scope, rigor, and fertility that the most reliable and productive efforts of contemporary science would be faulted without its governance.

Science is maximally reliable because it has at its disposal two progressively refined techniques which provide independent checks upon its claims: 1) observations which elicit involuntary recognitions irrespective of the commitment of the men, individual or collective, involved in inquiry, and 2) tests of internal consistency and licit derivation provided by the rigorous techniques of formal logic. These are the intrinsic devices of self-correction which make science something more than an ideology. Any ideology possessed of the independent techniques of self-correction, cognizant of the corrigibility of its descriptive or substantive claims, and subject to the consistency criteria of formal logic, is no longer an ideology in the generally accepted meaning of the term.

What has been argued is that science is a unique cognitive activity; it is the most reliable method for warranting both empirical and formal truth claims. The necessary conditions involved in establishing maximal reliability invoke direct or indirect appeal to unproblematic assertions about the nonlinguistic object world and tests of logical consistency and licit derivation. Science has no essential content, but is composed, at any specific period, of the collection of assertions’that individually enjoy maximal credibility and collectively have the greatest measure of internal and mutually supportive strength.

Against this account there has arisen, in the recent past, a counter thesis that sees science as only one of several “ideological” devices for producing conformity in the young, suppressing dissidence within the knowledge enterprise, and licensing the status quo. Science is not seen as a technique for providing responsible truth ascription, but as a device which is in and of itself essentially nonrational—a device calculated to insure the sovereign privilege of one or another group or groups. Science, in effect, does nothing to certify the truth of propositions, science assigns truth because of extraneous and noncognitive commitments.15 Science is part of the “Establishment’s” calculated efforts to dominate men’s minds.

Such a thesis has won considerable support among the disaffected and the professionally disadvantaged. As a thesis it is paradoxical at best. Its central argument is that science cannot make reliable truth ascriptions, that science is, in a radical sense, subjective, unreliable, serving special interests, and that its truths are so characterized only because they serve those special interests. And yet, for all that, those who advance the thesis have themselves discovered the truth that science cannot discover truth.

Science and Objective Truth

The question is, if science cannot certify truth, what certifies the truth of the thesis that science cannot certify truth? Is there a method independent of science that provides more reliable truth ascriptions? If so then that method is science and the objects of criticism are but the semblance of science. If the thesis is that there is no method for responsibly assigning truth status to propositions, the argument is self-stultifying, since the thesis itself is composed of propositions which individually and collectively claim truth status. One simply cannot have it both ways—but whichever way one chooses to have it, science is restored as the only method available for responsibly assigning maximally reliable truth status.

One way out of the obvious dilemma is to opt for a specially gifted segment of the community that escapes interest-bound “ideology.” This is, of course, the escape recommended by Karl Mannheim. Truth is attainable, but it is the privilege of a small community of declassed intellectuals. This “relatively classless stratum” has the advantage of true objectivity. It can discover and license truth because it is relatively untrammelled by prevailing interests.

When the thesis is advanced in this fashion, it simply means that science requires a specific and specifiable technique that would invariably identify instances of individual and collective bias in any pursuit of responsible knowledge about the natural and social world. The evidence of bias might manifest itself, for instance, in the neglect of anomalies unexplained or unattended by research scientists and/or scholars, or by flawed logic in the articulation of arguments. In effect, what such a thesis ultimately culminates in is an admonition that the methods of science should be applied more rigorously than has been the case to date—an admonition one can only applaud. We can only identify truth, whatever its source, if we have independent standards which characterize it. That the “relatively classless stratum” is best equipped with such standards and is thus more capable of identifying truth is itself an empirical claim—to be certified by the techniques of standard science which certify any descriptive claim.

An alternate, recent, and well-known attempt to throw into question the entire concept of the objectivity of science has been that undertaken by Thomas S. Kuhn. His challenging monograph, The Structure of Scientific Revolutions, has been employed by some to discredit any effort to attain objective truth in the field of social and political inquiry. It is at least doubtful that whatever Kuhn says about the natural sciences is transferable without supplementary and special arguments to political inquiry, but for the sake of discussion such a transfer can be countenanced.

Kuhn begins his argument by suggesting that there is something called a “paradigm” which governs activities which he, in turn, calls “normal science.” He seems to conceive “paradigms” as involving something like a special and specific research tradition embracing “law, theory, application, and instrumentation together” (Kuhn, Structure of Scientific Revolutions [Chicago: University of Chicago, 1962], p. 10). In “preparadigm” periods a science characteristically involves only random research and ad hoc scholarly efforts. During its “preparadigm” stages the pursuit of knowledge in a specific universe of inquiry is unsystematic and relatively unstructured. A “paradigm” develops by some “inscrutable” happenstance and constitutes the “promise of success” in some selected and incomplete ranges of concern (pp. 23f., 89). On the appearance of a “paradigm” a science is transmogrified. Once the “paradigm” becomes the dominant research orientation, “normal science” becomes a “puzzle-solving activity,” an activity that fulfills the promise, and is in the service, of the paradigm. The paradigm, Kuhn seems to argue in some places, provides not only the “puzzles” to be solved, but the rules and criteria governing the scientific enterprise itself (p. 42). Science is conceived to be, in some sense, intrinsically paradigm dependent. A paradigm comes to dominate a universe of inquiry by a process that Kuhn seems to feel is remarkably like religious conversion. The scientist gives his “allegiance” to the prevailing or a contending paradigm (p. 150) and to opt for one paradigm rather than another is to make a “transition between incommensurable,” like opting for Methodism rather than Roman Catholicism. When paradigms compete for “allegiance,” there seems to be “no point at which” resistance to one or the other paradigm becomes “illogical or unscientific” (p. 158). A decision to opt for a contending paradigm “can only be made on faith” (p. 157). The decision to transfer allegiance from one paradigm to another does not involve matters “of proof or error”—”the transfer of allegiance. . . is a conversion experience that cannot be forced” (p. 150).

This kind of reading can be given, unfortunately, to Kuhn’s interesting monograph—and there has been a disposition, among some political scientists, to so read it. As a matter of fact a more generous interpretation of Kuhn’s account can be provided and, once some of the more hyperbolic utterances Kuhn permits to escape are deflated, his narrative collapses into a variation (still interesting and important) of the standard account of scientific procedures.

First of all, it is not absolutely clear what Kuhn understands by a paradigm. On some occasions in referring to “paradigms,” he simply talks of “conceptual schemata.”17 “Paradigms” obviously perform all the functions that we shall see can be attributed to a variety of linguistic entities which can be collectively referred to as “conceptual schemata.” Conceptual schemata, as we shall argue, perform all the functions Kuhn attributes to his “paradigms”:

1) they provide a schematic guide to research because they are composed of 2) broad theoretical or speculative hypotheses about the object world. Because they characterize the object world in a broad but relatively distinct fashion, they 3) provide the occasion for determining what will count as significant data of observation and, finally, they 4) afford the categories in which relevant data are classified, catalogued, stored, and retrieved. But we shall argue that no one linguistic entity performs all these tasks. The schemata involved are many and they all differ in character and degree of sophistication. Kuhn seems disposed to telescope all these schemata under the generic rubric “paradigm.” But it is clear that the various linguistic entities housed under this rubric are substantially heterogeneous. On one occasion, for instance, Kuhn speaks of a paradigm “taken from some fraction of everyday discourse” (p. 126). This could hardly be a paradigm that could guide the specific “puzzle-solving” research of the most unsophisticated “normal science.”

We shall refer to such conceptual schemata as “ordinary language schemata.” They are the artifacts of ordinary language and are beset by all its linguistic vagueness and ambiguity. Only when a scientist shifts to a more sophisticated linguistic level can such a schema begin to function as a “framework,” or a “conceptual model,” that provides an initial definition of his field of inquiry. Such “frameworks” shall be referred to in the subsequent discussion as “preliminary conceptual schemata” — they suggest speculative hypotheses and begin to isolate certain elements in experience as significant. Certain variables are understood to have causal relevance in the analysis of experience. In the effort to provide evidence for the hypotheses suggested by these preliminary conceptual schemata, a social scientist will be compelled to specify the semantic meaning of critical terms in his hypotheses and reduce the syntactical vagueness of his formulations. As we shall suggest, the result is partial standardization and formalization. What subsequently develops is a partially or fully axiomatized system. In some cases a calculus develops and, once conjoined with coordinating definitions which relate the calculus to the object world, we have a partially or fully articulated standard theory. Such a theory can operate as a “paradigm.” Such “paradigms” are to be found in the formalized sciences like physics and chemistry. Political science simply has no “paradigm” of this kind (Kuhn seems to recognize this; ci. p. 15 of his text). Political scientists are not involved in “puzzle-solving,” because “puzzle-solving” implies, as is clear in Kuhn’s account, a viable theory covering at least a substantial range of phenomena within a specific universe of inquiry. Political science has no “general theory” of politics that could pass as a “paradigm.” At best, even by Kuhn’s account, political science is “preparadigmatic.”

But more interesting for our purposes here is the analysis of his account of “paradigm-switching” which could be construed as “subjectivistic”—denying to standard science the objectivity which makes it what it is.

If the account of conceptual schemata offered in the foregoing account is correct, then “rules,” “standards,” and “criteria” which establish the truth of the propositional constituents of a paradigm (or theory) can hardly be by-products of the paradigm itself. Paradigms develop because scholars and research scientists pursue the methods of science . . . and those methods are “transparadigmatic.” The methods produce paradigms; it is not paradigms that produce methods, rules, standards, or criteria.

Kuhn seems to recognize as much. He regularly refers to “anomalies” which cannot be “explained” within a prevailing paradigm (pp. 6, 33f., 52, 56f., 62,65, 74f., 8 if., 86). These irrepressible anomalies are intersubjective and unproblematic observations recognized by practitioners “within” the paradigm—they persist no matter how strong the “allegiance” to the established paradigm. In effect, the acceptance of a paradigm is contingent—the measure of its acceptability is a function of its ability to explain and predict the phenomena recognized in critical ordinary observations which must of necessity, be “extraparadigmatic.” With the increment in the number of anomalies unexplained by the regnant paradigm, scientists begin to contemplate alternate ways of structuring, in theory, their experience. An alternate paradigm, which is actually an alternate preliminary conceptual schema, is advanced because it is conceived as offering solutions to the problems unresolved by the prevailing paradigm. It is neither “accepted,” nor does one give “allegiance” to it. It is advanced as a candidate (as yet informal) theoretical sketch. To be accepted it would have to go through the stages which we shall characterize more fully in the forthcoming account. Its speculative and loosely worded hypotheses would have to be reformulated into test hypotheses and evidence accumulated that would satisfy intersubjective tests of adequacy. And those tests are not “paradigm dependent.”

The “observations” which are referred to as “protocol statements,” or “observation statements” do function, at least in part, within the compass of ordinary language, but they are not dependent upon the paradigm under scrutiny. The very fact that “anomalies” develop within the prevailing paradigm indicates that not all observations are “paradigm dependent”—if they were, such anomalies would never appear and every paradigm would be forever insulated against empirical disconfirmation. All of which means that paradigms simply cannot be conceived as sui generis and incommensurable. Their acceptability is governed by their ability to more effectively explain and predict publicly observable phenomena. Kuhn, himself, says as much in a number of places. A new paradigm does succeed in explaining the anomalies that plagued the old paradigm—and predicts those unsuspected by the antecedent paradigm (pp. 1 52f.). The new paradigm is more fruitful in opening up new areas for detailed inquiry (pp. 10, 103). This indicates that the rules governing empirical evidence, explanation, and prediction are “transparadigmatic”—they are the rules of evidence, explanation, and prediction governing standard science—and they are substantially independent of any paradigm.

At various places in his treatment Kuhn, himself, further indicates that rules of evidence and the criteria governing consistency and logical derivation are “paradigm independent.” He indicates, in fact, that paradigms are measured against logical criteria of consistency and simplicity (pp. 153f.). All of which indicates, if it indicates anything, that logic and empirical evidence criteria subtend science in general and are not paradigm specific.

What Kuhn seems to be arguing is that an initial conceptual schema cannot meet the test of evidence and yet is advanced on the supposition that it will resolve the logical and observational perplexities which infect an antecedent paradigm. That such a preliminary schema is offered and acted upon, he describes, unhappily, in terms of acting “on faith.” But such an “act of faith” is similar to an “intuition” or an “insight,” which suggests a particular manner of conceiving problems. No “intuition” is born in “faith.” It is the result of long familiarity with the procedures of standard science.’8 “Intuitions” are elaborate “hunches,” guided largely by what one has learned about a particular universe of inquiry. It is true that each candidate preliminary schema is offered in the hope that it will facilitate the resolution of problems, provide the schematic guide for research, suggest test hypotheses, afford taxonomies and classificatory schemata serviceable to the identification, storage, and retrieval of information, and constitute a mnemonic convenience to the cataloguing of information. But a preliminary conceptual schema, one logical level removed from the schemata that inhabit ordinary language, cannot hope to “prove” its utility. Only a fully articulated theory can do that.

To call all schemata “paradigms” is to obscure the many and significant differences between a variety of linguistic artifacts each having different evidence conditions and theoretical yield. When a fully articulated theory meets the evidence criteria of standard science, it is unreasonable not to accept it. And this is precisely what Kuhn says (pp. 30, 146, 157). He says that the Newtonian paradigm was finally accepted by the scientists of the eighteenth century—and that they “had every reason to do so” (p. 30). Similarly, in his volume on the Copernican revolution, Kuhn argues that the Ancients refused to countenance the heliocentric conception of the universe advanced by Aristarchus and, because of its prima Jade implausibility, the lack of confirming evidence, its inability to predict phenomena, and its logical peccability, they had “every reason to do so.”19 If men have “every reason” to accept or reject alternate paradigms, one wonders what “conversion” or “opting on faith” have to do with the cognitive process of discriminating between alternative paradigms. The fact is that the evidence requirements for accepting a preliminary conceptual schema are different than those for entertaining a formal theory. Newton offered a reasonably well-articulated theory; Aristarchus offered, at best, a preliminary conceptual schema. The fact that the highly modified analogue of Aristarchus’ heliocentric conception ultimately proved acceptable to the scientific community means that “brilliant insights,” “intuitions” and “hunches” may be true—just as the evolutionary notions of pre-Socratic and Lucretian speculation have proved more akin to contemporary conceptions of the evolution of species that the Aristotelian and Linnean concepts of persistence of species. But a clear distinction must be made between preliminary conceptual schemata, often brilliant and suggestive, but too vague and general to be confirmed or disconfirmed—and extensively for- malized theories which are subject to logical and empirical truth conditions. To call all the linguistic entities advanced by thinkers “paradigms” is to obscure fundamental differences between them. Where no distinctions are made—anything can be said . . . unfortunately with little cognitive profit.

One can admit therefore almost everything Kuhn says and yet not submit to his notion that paradigms cannot be subject to “transparadigmatic” evidence conditions. As has been suggested, the fact that anomalies develop within prevailing paradigms is clear evidence that observations are made with relative independence of them. Furthermore, that paradigms are subjected to logical scrutiny indicates that not only critical observations, but logical procedures, have an independence of paradigms.

It is clearly the case that “faith,” “interests,” and “parascientific commitments” figure in the acceptance and rejection of alternate conceptions of man and society. This is an interesting and significant sociological and psychological fact. But such considerations have nothing whatever to do with the analysis of the admissibility conditions governing truth ascriptions. The fact that men permit non- cognitive concerns to influence their judgments is a well-known and generally acknowledged fact; but that it is, in principle, impossible to isolate those influences is a claim that has not been established. A distinction must be recognized between per/ormance and competence. Many natural and social scientists, indeed whole political cultures, admit “truths” only if they conform to some religious, political, or philosophical prejudice. But the fact that men so perform does not constitute evidence that they can only so perform.2° Kuhn’s indisposition to make the necessary distinction between the logic of discovery—how men have tended, in fact, to argue about alternate interpretations of nature and man—and the normative and prescriptive logic of justification thwarts analysis and confuses issues (cf. pp. 8f.).

We accept a proposition, or an articulated collection of propositions, not on impulse, intuition, or faith—if we pretend to be rational—but rather because it fits coherently within a system of funded credibilities, themselves warranted, individually and collectively, directly and indirectly, by best evidence. Best evidence for empirical claims refers to assertions about matters of fact that survive the scrutiny of intersubjective assessment. These are the “protocol utterances” of natural language. The import of such utterances is always directly or indirectly referential. We anticipate futures through them. We utter them because some experiential cues tell us that we have an instance of some identifiable class of objects before us. If our anticipations concerning them are fulfilled, we say our utterance was confirmed. If we fail to satisfy anticipations, we review the system of credibilities in which that utterance is lodged and which contributed to its articulation. The ultimate purpose of all this activity is to permit us greater reliable leverage on life and experience. We understand because we can predict and explain. We can predict and explain because we can make significant and recurrent discriminations in practice. We guide this activity with formal and informal logic and inductive procedures. The entire program is that of standard science.

Science, Ethics, and Political Science

There is no intrinsic reason why this program cannot include ethical concerns. Matters of ethical moment, for example, can be studied as the subject matter of empirical science. Ethical utterances can become the first-order constituents of a second-order scrutiny (when we analyze ethical language in metaethics), and we can seek out the meaning and implications of ethical “principles” using standard techniques of analysis, descriptive and probability assessment. All these activities are eminently cognitive—and it is doubtful if there are any human concerns which are irreducibly and forever outside their ken.

Political inquiry is an informal discipline. Much of its research and scholarly activity is fragmentary and seemingly random. Individuals and groups of individuals pursue some special concern with the techniques suitable to their purpose. There are those preoccupied with axiomatizing some or another area of inquiry— and their preoccupations are essentially analytic. There are those who are concerned with how participants in the political processes actually do behave, or they are concerned with how they might behave under certain initial conditions. Their activity concerns itself with adequate description and, more frequently than not, low-level prediction. Their concerns are essentially descriptive. There are those who focus on the values men harbor or pretend to harbor. They seek out the vindications that men might offer to support one or another vision of the world and of men and their society. They are concerned with the issuance and support of normative utterances. This is perhaps one of the least understood areas of political concern. But the efforts that characterize this activity are essentially cognitive and science, broadly conceived, constitutes its substance.

To have characterized these distinct domains of concern is not to suggest that they are mutually exclusive even if exhaustive of the preoccupations that animate political inquiry. There is a complex and important interaction between these analytically discrete domains. Very frequently those concerned with analytic preoccupations hold naive and irresponsible normative commitments—and succeed in drawing, as a consequence, illicit conclusions from their inquiries. Similarly, and more frequently, this is the case with empirical political scientists who claim that their undertakings are “value free” in some vague and unspecified sense. On the other hand, many political theorists, who claim special competence in the normative domain of inquiry, simply lapse into confusion and conjure up a new obscurantism which reduces serious concern for political issues to a contest between “ultimate values” which are the gift of “faith,” “phantasy” and/or the “dialectic.”

Political science is a coherent and unified concern with political life. Because it is informal, minimally standardized, and possessed of only a restricted inventory of confirmed generalizations and can exhibit no general theory which covers its universe of discourse, its inquiries, for the foreseeable future, will be defined in a variety of ways and its pursuits will be, in large measure, significatly influenced by the personal concerns of individual scholars and research personnel. But subtending all this activity are implicit rules governing truth ascription—rules which make the activity coherent and its results cumulative. The task of making those rules explicit is arduous, sometimes thankless, and in a substantial sense endless.

Those rules originate in the life process itself—and life is an ongoing activity. In attempting to characterize that protean activity we employ language and we do not, and cannot, have an ideally antiseptic language that fails to give rise to confusions and creates no paradoxes. But this does not constitute an argument that because natural language is vague and the source of confusion and paradox— vagueness, confusion, and paradox are to be embraced as truths truer than true.

Political science is, at best, a partially formalized and standardized science, a corrigible and self-corrective pursuit of propositions of maximal reliability concerning political matters (however defined). It is a pursuit governed by the reasonably well understood domain invariant criteria of logic and the domain variant criteria ofempirical research, theory construction, and ethics. Its successes, by whatever standards, have been relatively modest—but whatever successes it has enjoyed have been achieved through the use of standard science broadly conceived. Neither the addiction to linguistic magic nor the advent of a new obscurantism can confound that fact. We do know a great deal more about man’s political life and his political behavior today than we did a generation ago. Our future success will be contingent on our ability to effectively employ the corrigible methods of contemporary science, on our ability to refine those methods, and on our capacity to more cognitively assess our needs, aspirations, and conflicting desires as human beings. There are no magic formulas nor guarantees of success. There is only the prospect of hard and collaborative enterprise.

Science, Political Inquiry, and Metapolitics

If there is any merit in the preceding discussion, it trafficks on an analysis of several key concepts: “science,” “objectivity,” and “conceptual schemata” among the most critical. It is equally evident that the issues involved and the concepts employed require far more substantial treatment if the analysis is to be convincing. The point of this introductory essay into the issues that seem to invoke so much contemporary energy is to convey an appreciation of the fact that before the student of political inquiry can address himself to the “significant” concerns that beset the discipline, it is necessary to gain a foothold on the rudiments of cognitive activity itself. In order to accomplish that, it is necessary to unpack terms in common currency—terms whose meaning seems to be intuitively understood but whose meaning seems to become more variable in direct proportion to the frequency of their use.

There is, in fact, a catalogue of terms that the student of political inquiry employs, each of which has a variable, vague, and ambiguous meaning. The term “meaning,” itself, has variable, vague, and ambiguous meaning. Similarly, terms like “truth,” “law,” “theory,” “explanation,” “model,” “understanding,” “knowing,” “normative,” and “noncognitive,” have elusive and fugitive meanings. Before meaningful discussion can take place between participants in the enterprise we identify as political inquiry, some initial moves must be made to reduce the variance that characterizes too much of our talk.

What this means is that linguistic analysis is a necessary preliminary to political inquiry if we are to attempt any adequate assessment of the diffuse and complex issues that agitate political inquiry. If we are to more competently pursue any of the issues joined in this introductory discussion, we shall have to address ourselves to the conceptual language with which such a discussion must be conducted. In effect we shall employ a metalanguage to discuss the language of political inquiry. Hopefully what will result will be insight into the implicit and explicit rules governing man’s most distinctive and successful cognitive activities. It will be, at best, a foray into a vast and poorly charted territory—a brief introduction to metapolitics. [2]

A. James, Gregor, “Metascience & Politics, An Inquiry into the Conceptual Language of Political Science,” 2nd, ed. (New Brunswick, New Jersey: Transaction Publishers, 2003).

[1] Gregor, A. James, Metascience & Politics, An Inquiry into the Conceptual Language of Political Science, 2nd, ed. (New Brunswick, New Jersey: Transaction Publishers, 2003), pp. 123-140.

[2] Gregor, A. James, Metascience & Politics, An Inquiry into the Conceptual Language of Political Science, 2nd, ed. (New Brunswick, New Jersey: Transaction Publishers, 2003), pp. 17-38.