20th World Congress of Philosophy Logo

Philosophy of Mind

The Concept of Intelligence

Ira Altman
City University of New York

bluered.gif (1041 bytes)

ABSTRACT: Gilbert Ryle’s dispositional analysis of the concept of intelligence makes the error of assimilating intelligence to the category of dispositional or semi-dispositional concepts. Far from being a dispositional concept, intelligence is an episodic concept that refers neither to dispositions nor to ‘knowing how,’ but to a fashion or style of proceeding whose significance is adverbial. Being derivative from the function of the adverb ‘intelligently,’ the concept of intelligence does not have essential reference to specific verbs but rather to the manner or style of proceeding of nearly any verb that is descriptive of the proceedings of an agent. Intelligence- words are expressive of a manner of doing things that may be narrated in one of two ways. The first takes the form of a series of contrasts which, when put together as a list of disjuncts, may be called the contrast-criteria of intelligence. The second may take the form of the characteristic activities which comprise the criteria of intelligence.

bluered.gif (1041 bytes)

This subject is but a small part of the larger issue that is waged between dualists and materialists: whether the words used to ascribe mental qualities have a physical or "psychological" reference. Much of the literature concerned with this broader topic takes on the character of a general broadside against either the materialist or dualist position. When specific qualities of our mental life are discussed, it is discussed either in passing, or to make certain that they be assimiable to the general thesis being propounded. When Gilbert Ryle wrote of intelligence,(1) he was interested in making it out to be a dispositional concept (his technique for refuting dualism). Ryle's fear seemed to have been that if any mental concepts escape such dispositional analysis, then we may be forced to give way to dualism. The dualist's reaction was to withdraw from the dispositional analysis and to claim that " intelligence " can only make sense when it is ultimately linked to an alternative concept which presumably has reference to a "process" (the existence of which may be ascertained by a subjective avowal: intent, deliberation, reflective choice, etc.).

Whether this maneuver avoids the Rylean critique is subject to serious doubt. Each new term may itself have a problem concerning the "metaphysical" status of its referent: whether it is a physical or a "spiritual" process.

Another problem with this approach is that it often makes the concept of intelligence dependent on factors that may have no more than a casual relation to its meaning: a contextual connection, for example.

While a dispositional analysis of intelligence would effectively rescue it from entrapment inside "the ghost in the machine" it would fail to do the concept the justice that the dualist seeks for it, and that is to make the observation that it appears to be more an episodic than a dispositional concept.

It is in this connection that some philosophers take a critical view of Ryle's analysis. Peter Thomas Geach, for instance, claims that reports of mental acts are categorical and are not hypothetical or semi-hypothetical statements about overt behavior. He believes the dispositional account of psychological concepts to be on a level with the statement that opium puts people to sleep because it has a dormative power, and that such an account if it were carried over to the physical sciences would be met with impatience.

A physicist would be merely impatient if somebody said to him: "Why look for, or postulate, an actual difference between an unmagnetized or a magnetized bit of iron? Why not say that if certain things are done to a bit of iron certain hypotheticals become true of it?" He would still be more impatient at being told that his enquiries were vitiated by the logical mistake of treating " x is magnetized " as categorical, whereas it is really hypothetical or semi-hypothetical.(2)

This is almost certainly a misrepresentation of the Rylean position. Ryle's analysis is in keeping with the Wittgensteinian thesis that mental language has no private sense, and while Ryle uses this thesis to undercut an ontology of mental events (a procedure resisted by Wittgenstein), his position is not (as Geach would have us believe) incompatible with the occurrence of physical processes that are explanatory of the hypotheticals that constitute a part of our mental conduct language.

To say that the brittleness of glass (or the magnetism of a bit of iron) is a dispositional concept is not to deny that the hypotheticals in which that concept is couched may not be explicable in terms of categoricals such as a molecular substructure, or stress and force factors. The hostility to Ryle's analysis comes less from a scientist's impatience with hypotheticals, and more from a metaphysician's perception that Ryle's use of the term "categorical" is sometimes a euphemism for "private, non-physical event." The dualist then makes use of this terminology with a special vengeance: "If categorical uses of mental-conduct concepts exist, then mental events must exist." But of course we should be careful here, for this does not necessarily imply that categorical mental statements are statements about nonphysical (ghostly) processes.

An earlier and more cautiously drawn critique of Ryle occurs in Stuart review of The Concept of Mind. He poses the question: "On what grounds does Professor Ryle decide that there are no acts answering to such verbs as 'see,' 'hear,' 'taste,' 'deduce,' and 'recall'...?"(3)

He concludes that Ryle has mistakenly identified the meaning of a statement with the method of its verification:

Because overt behavior often constitutes for most people the best and, in some conditions of utterance, the sole available evidence for statements about mental activities and states of minds, such statements can be idenified with hypothetical statements about behavior.(4)

Ryle's position is likely to prove too elusive for this criticism. While it is true that it looks as though Ryle's program is to re-interpret categorical statements about "mental events" as hypothetical statements about publicly observable events, his position is not sharply distinguishable from Wittgenstein's thesis about mental terms. Wittgenstein did not deny that people have private thoughts or that they have feelings they do not show, and neither does Ryle. Wittgenstein did want to deny that mental talk could have a meaning independent of some connection with a public language that posesses public rules of discourse. After all, Ryle does say (as even Hampshire recognizes) that for any given statement there are an indefinite number of possible answers to the questions: "How do you know?" or "What are the reasons for saying?"depending on who made the statement, when and in what circumstances. Though Ryle may be less than clear about the general thrust of his thesis, it may be too strong for us to accuse him of covert verificationism. Ryle makes too many hedges against reductionism to be providing us with a "general account of the logic of statements ... where 'the logic' means 'the method of establishing the truth of...'"(5) However accurate (or inaccurate) these interpretations of Ryle's analysis may be, it is important for us to distinguish our position from Ryle's and show whether the aforementioned critiques have any relevance for our thesis. I am in agreement with those critics of Ryle who insist that mental terms often (even usually) have categorical and, therefore, episodic uses. In this respect my thesis is untouched by the criticism that Ryle's dispositional analysis attempts to eliminate the occurrence uses of mental concepts.

Mental terms have episodic uses, and "intelligence," derives virtually all of its ignificance from such a use. But I do not agree that an occurrence or episodic analysis of mental concepts opens a way toward the re-introduction of mental events as "ghostly processes." Insofar as the concept of intelligence is concerned, the occurrences referred to are publicly observable: the significance of " intelligence " is predicated on outward criteria. My analysis is intended to be a corrective of the Rylean thesis by acknowledging some of the claims of his critics but without at the same time accepting the metaphysical dualism that such an admission is believed to yield.

We can readily see, from every-day discourse, that a large percentage of words connected with intelligence turn out to be adverbs. This is not unusual since so much of what we say about intelligence is not simply that so-and-so, or this-and-that is intelligent (that someone generally does something, or that something is done) but that it is done in a special way—whatever the verb is, it is modified. Adverbs generally have limited ranges of application. It does not make sense to attach them to any verbs whatsoever.(6) It will be important to determine why some actions and the verbs describing those actions may be modified by the adverb intelligently, and others not. Such an analysis should yield a group of verbal narratives informative of the mechanics, as well as the "conditions" that make the use of the concept of intelligence acceptable or unacceptable in any given context.(7) The appropriate narrative descriptions (such as insight, flexibility, and novelty) will be familiar to anyone who speaks the language. What may be surprising is the nature of the conceptual link between such narrations and "intelligence." The "logical status" of these descriptions as "conditions" for the application of the terms "intelligent" or "intelligently " may be neither "analytic" nor merely "synthetic.

The variety of narrative descriptions that the adverb "intelligently" collects may be parcelled out into at least two different sets of "conditions" for the application of the term, each with special and different contexts together with different sets of "rules" for that application. Some concepts which were thought to be somehow "intrinsically" linked with intelligence (such as "intention" or "learning") will be found to be either a part of the general context (not a condition) of application, or to have only a contingent relationship to that concept. The result is an analysis of intelligence that is neither dualistic nor merely dispositional, but one which hopefully captures the varieties of uses made of the concept. The general thrust of my thesis is to support the Rylean analysis but to also show that limits exist between the application of the concept of intelligence, and other mental-conduct concepts.(8)

It will be necessary to demonstrate a distinct difference between the concepts of ability, "know-how", and intelligence (as a failing in the dispositional analysis). Also, we need to demonstrate the differences that exist between such concepts as intention, purpose, learning, and intelligence, thereby blocking their purported candidacy as analytic conditions for the application of intelligence terms. Examples can be chosen from the psychological and biological literature to illustrate both technical and commonplace uses of the words of intelligence in order to obtain a sampling that can serve as "criteria" for the application of the words of "intelligence." This will require a discussion of the status of such "criteria."

I believe that such criteria cannot be described as either analytic or synthetic, and that there does not exist a single criterion or single class of criteria, but rather two groups of mutually compatible criteria, each with a different basis of use, dependent on differences of comparison. In The Concept of Mind, Ryle makes the claim that " intelligence " is a " determinable " disposition and that it is a species of " knowing how," a kind of ability, competence, or skill. His argument for the first of these claims focuses on the dispositional use of the concept expressed by the adjective "intelligent," and makes the claim that in order for a concept to be determinate, its exemplary (action-contents) must invariably be found in the verb expresses its occurrence. Thus "baker" is a highly determinate dispositional concept because there exists a specific verb that is its exemplary: "to bake." Should a dispositional concept be lacking a verb, Ryle identifies it as highly generic or determinable: a grocer does many different kinds of things. There is no verb answering to an essential and specific activity of the grocer, there is no "grocing" activity, and this latter condition holds for the concept of intelligence as well, making it a highly generic or determinable concept.

Ryle's use of the determinate-determinable distinction leads to misconception due to an ambiguity in that distinction. There appear to be two separable notions in the idea of the "determinable:"

(1) a determinable concept (proper), and

(2) a polymorphous concept

To say that a concept is polymorphous is to say that there is not one unique kind of action or occurrence associated with it, but a wide range of different kinds of actions and occurrences. To say that a concept is determinable is to say that the totality of actions and occurrences associated with it have never been (nor are they likely to be) enumerated. When Ryle says that " intelligence " is determinable he means to say that it is both polymorphous and determinable. But while he provides evidence of its polymorphousness, he does not show it to be determinable. Showing that a concept is one of these is not showing that it is the other. A concept may be polymorphous and determinate, and Ryle has not shown that " intelligence " is not so.

Intelligence has an adverbial significance. Derivative from the function of the adverb, "intelligently, " as a verb modifier, the concept has essential reference not to specific verbs, but rather to the manner or style of proceeding of nearly any verbs descriptive of the proceedings of an agent, so that whether the concept of intelligence is determinable or determinate is not discoverable from grammatical features alone. The various contexts in which agents are said to act intelligently need to be surveyed in order to elicit the used for the application of the concept. Whether the criteria vary without end or are invariably repeated (or something in between) can only be uncovered by such context imbedded surveys.

Another distinction must be made for an avoidance of the Rylean error and that is the distinction betwen exemplaries and occasions. The verb in " Aristotle reasons logically " is not an exemplary of the tendency to be logical as the verb in " He baked a cake " is an exemplary of the tendency to bake. " Reasons " in this case is an occasion in which the tendency to be logical is exemplified by the manner of reasoning that Aristotle exhibits, but the verb by itself does not do the job of exemplification (if at all), it is the adverb "logically" that clues us in to the disposition-type or action-contents being referred to. The exemplaries of a concept are either analytically or " criteriologically " connected with the concept, while occasions need not be and usually are only contingently connected. Coupling an exemplary with the adverb of the concept leads to redundancy in a way in which the coupling of occasions with the adverb does not. While there is no redundancy in " Aristotle reasons logically, " there is redundancy in " Aristotle draws valid inferences logically."

By failing to distinguish between exemplaries and occasions Ryle erroneously assumes that the action-contents of a disposition (that is , its exemplaries) can only be expressed by its verb, otherwise the concept must be indeterminate. If there are no unique action-contents (as expressed by a verb) that serve as necessary and sufficient conditions (criteria) for the application of that concept, then it has no " essential " characterisitics. This it appears to Ryle is the case with " intelligent " for there is no unique verb of intelligence such as "intelligenting." However, contrary to this claim, intelligence words (when they express dispositions) are descriptive of the tendency to proceed in a special way, and are not descriptive of any of an innumerable variety of proceedings, no matter how innumerable the occasions in which that way of proceeding may be exemplified. Intelligence words find their exemplaries (action-contents) not in those occasions which are described by the verbs of sentence, but in the episode-qualifying narratives that the adverb of a sentence collects. To put this another way, the disposition to intelligence as expressed by the adjective 'intelligent' finds its exemplaries (action-contents) in those process-fashions or episode-manners that 'intelligently' (the adverb) names. There does not exist a unique verb of intelligence,but this does not mean that intelligence is determinable, it may well be polymorphous and determinate.

Ryle's claim that intelligence is a species of " knowing-how " runs afoul of the fact that abilities can be exercised unintelligently whereas liabilities may be exercised intelligently.

The position that intelligence and know-how are more or less the same thing lends credence to the supposition that intelligence is an indeterminate concept, for the number of things we may know how to do are indefinite, and so if intelligence is a kind of ability, then it too is indeterminate. But since intelligence is not an ability at all but rather a capacity (something that cannot be acquired) this support for the indeterminacy thesis fails.

The insistence that unless a concept has exemplaries that are both necessary and sufficient conditions of its application, then it is determinable, is overly-stringent. Wittgenstein has shown us how it is possible in ordinary language to obtain conditions for the application of a concept that are not so sharply drawn as is analyticity, but which nevertheless permits us to obtain a disjunction of criteria that are stronger than synthetic but weaker than analytic. This opens the way for the possibility that the characteristic components of the concept in question (even if it be polymorphous) form a closed list (that is the list of disjuncts are not indeterminate but finite).

J. Holoway's analysis of the concept of intelligence suggests that aside from criteria of intelligence, there may also be criteria that focus on the antitheses of intelligence, or on intelligence-contrasts, that would allow us to recognize the intelligence of specific acts in specific contexts. On this account, an act would be judged intelligent once such other possibilities as chance, instinct, and habit are ruled out of consideration. It appears then that, depending on context, the criteria of intelligence may sometimes " describe " action-contents (exemplaries) characteristic of intelligent behavior, or provide us with a denial that the action-contents in question are characteristic of instinct, or habit, or reflex, or chance, and so on.

Such a listing of disjuncts may be called contrast-criteria, for these are contrasts between different kinds of process-fashions or styles of proceeding. Both sets of criteria for intelligence operate within a context that this concept has in common with such concepts ashabits, tropisms, and instincts, and this is a means to ends context. They are all purpose-adapted styles of proceeding, and this may be one reason why some of these are confused for one another.

I conclude that intelligence-words are expressive of a manner of doing things that may be narrated in one of two ways. The first of these narrations takes the form of a series of contrasts, which when put together as a list of disjuncts may be called the contrast-criteria of intelligence:

i. If I, then—(C v R v T v H v N)

ii. If—(C v R v T v H v N), then I

where C is chance, R is reflex, T is tropism, H is habit, and N is instinct. Quite possibly R, T, H, and N might be reduced to A: automatism, in which case the contrast—criteria could be represented as:

iii. If I, then—(C v A)

iv. If—(C v A), then I

Another descriptive narration of intelligence may take the form not of contrasts but of characteristic activities, the criteria of intelligence:

v. If I, then (Fs v Fn v M v T v P)

vi If (Fs v Fn v M v T v P), then I

Where Fs is self-corrective flexibility, Fn is inventive flexibility, M is multiplexity of routines, T is transference of acquired routines and P is " a sudden change in the learning curve " or (less technically) probable success.

bluered.gif (1041 bytes)


(1) Ryle, 1949.

(2) Geach, (1956, 62).

(3) Hampshire, (1949-1950, 242).

(4) Ibid., 247

(5) Ibid.

(6) If such limits of application exist, they would constitute a basis for the rejection of Ryle's claim that the dispositional concept of intelligence is determinable rather than determinate.

(7) "Not merely do adverbial expressions pick out classes of action, they also pick out the internal details of the machinery of doing actions, or the departments into which the business of doing actions is organized."Austin, (1970, 193).

(8) It is because of a failure to set limits (as in Ryle's description of "intelligent" as a determinable disposition) that the concept of intelligence becomes so easily confused with such concepts as learning and "know-how" in the minds of some philosophers (and I may add, psychologists) to the extent that they fail to see some rather obvious differences that exist between them.

bluered.gif (1041 bytes)


Back to the Top

20th World Congress of Philosophy Logo

Paideia logo design by Janet L. Olson.
All Rights Reserved


Back to the WCP Homepage