General and Special Abilities

ABILITIES, GENERAL AND SPECIAL. The now common term "general ability" in its current signification is of recent origin. Previously it appears to have had only occasional usage in ordinary writing and conversation, being there taken to characterize excellence in performances of several different kinds; especially when these were not all readily subsumable under the name of "intelligence." As for the professedly psychological treatises, they took little notice of individual differences of ability at all, and when they referred to such they commonly did so in terms of the orthodox faculties, such as acuteness of sense, pro fundity of intellect, tenaciousness of memory, liveliness of imagi nation, and so forth.

But in the year 1904 the observation was made by C. Spearman that the correlations between different abilities tended to bear a special and regular relation to each other. They showed such comparative magnitudes that, when set out in the usual square table, they could be arranged in a "hierarchical" manner, so that the values proportionately diminished from left to right and from top to bottom. And from this hierarchy mathematical reason ing led to the theorem that the measure of any individual in any ability could be divided into two factors: the one was common to all the abilities concerned, whereas the other was specific to that ability alone. These two factors, "general ability" and "spe cific ability," were designated by the letters g and s. Further, the important corollary was drawn that, by averaging together many dissimilar mental performances, the influence of the s's involved would—since they were mutually independent—tend to cancel out ; in the long run the influence of g, being constant, would be come paramount, or, in other words, the average would become an accurate measure of the pure "general ability" itself.

A year later, such a hotchpot of diversified mental performances was actually tried out, viz., by A. Binet, his purpose being to com pare the mental powers of exceptionally dull children with those of normal ones. But he avoided using the theory by which such a hotchpot had been suggested ; for this theory was incompatible with the doctrine of "faculties" which up to that time had been advocated by him and almost everybody else. Instead, he said that he was measuring what he called the child's "level of intelli gence." This he expressed in terms of age. If a dull child of x years could pass the test-performances about as well on the whole as a normal child of y years, this y was taken to represent the "mental age" of the dull child ; x minus y was called his "retarda tion." The success of these tests of "intelligence"—constructed, as they were, with extreme ingenuity—proved to be quick and great. School teachers took them up not only for the dull chil dren but for all others, however bright. Each child had his "men tal age" determined. So used and interpreted these tests found widespread favour. Other sets of tests were devised, and these proved to be no less successful, although possessing little, if any thing, in common with the original set, except that they were miscellaneous.

The next landmark in the movement was a surprising exploit. Several different tests of "intelligence" were applied to the Ameri can army, that is to say, nearly two million men. The results were declared to be highly satisfactory by the secretary of State for war, himself. Even this application of the tests, unsurpassable in size as a single event, was soon overtopped by their cumulative further employment elsewhere. Putting together all the indi viduals tested in various countries—school children, university students, industrial employees, and others—the sum total could hardly have been less than some twenty millions. Still the results were usually reported to be very successful. Teachers and other persons using them declared that they derived valuable informa tion from so doing.

Criticism.

Nevertheless, all was not quite so harmonious as it seemed. From the very beginning there had been an opposi tion party, and this continually waxed in strength. Many of the objections raised, indeed, were of but little importance. But one of them became increasingly obvious and menacing. It was that the testers did not really know what they were testing, nor even whether they were all testing one and the same thing. They were calling it "intelligence," but their usage of this term was highly equivocal. Few had troubled themselves to define -what they meant, and those doing so had disclosed the most discrepant views. Equally discrepant, too, had been the test-making in actual prac tice. Consider, for instance—the objectors urged—that large part of mental activity which is commonly entitled memory. Some of the test-makers had brought this within the pale of what they were naming intelligence, whereas others had as carefully kept it all outside. Even the same psychologist would on one occasion construct his set of tests so as to involve much memory and even formal memorization, but on another occasion would reject any such thing. As for Binet himself, he began by declaring memory to be not intelligence but only a mimicry of this, and yet he straightway proceeded to make his tests depend on memory in a very high degree. Just as unfortunate had been the state of affairs with respect to the faculty of imagination, for this also was in cluded by some testers within the purview of intelligence as con ceived by them, whereas by others it was explicitly excluded. But the climax was reached when several declared that under intelli gence must be comprehended even "sensation." Thus Binet him self wrote, "A sensation, a perception, are intellectual manifesta tions as much as reasoning is." Whereas other testers declared that if ever, from Plato onwards, there has been any one char acteristic to form the very core—the raison d'être—of the con cept of intelligence, it has been that of being something to be dis tinguished from and opposed to mere sensation. Small wonder, then, if the critics proceeded to assert, as J. Wallin did, that these widely-employed sets of tests of the so-called intelligence really "measure qualities which are so different as to be prac tically incommensurable." There was a further and even graver objection raised against these measurements of intelligence or general ability, or what- ' ever we like to call it. It demands evidence that any such mental power is susceptible of measurement at all. It points out that Binet had written as follows : "The mental faculties of each sub ject are independent and unequal; with a little memory there may be associated much judgment. . . . Our mental tests, always special in their scope, are each appropriate to the analysis of a single faculty." If thus, it is urged, the faculties are not one but many, how can one single measure possibly suffice? On such grounds, G. Whipple and S. Colvin had protested that a high total score in any such set of tests "fails to get anywhere in our real enquiry as to just which ones of the various mental functions are possessed by the extraordinarily-heightened efficiency. Is it memory span, or capacity for concentrated attention, or ability to handle symbols, or apprehension of abstract relations, or acute perceptive capacity, or lively imagination, or originality, or breadth of associative tendencies, or speed of learning, or what, that demarcates such a child from other children?" The inference from all this might seem to be that each of these faculties or functions ought to be measured separately. But thereupon just the same objections are raised over again, only on a smaller scale. For how, it is asked, can anyone measure memory in general when the individual may be excellently en dowed with it for some classes of objects but very badly for others? As much may be said for all the rest of the faculties, "judgment," "imagination," "attention," etc. There would ap pear to be no rallying place along this line until we reach the last elements of knowledge and agree with the Herbartians that every idea has its own imagination, its own memory, its own judgment, and so forth. In like strain Thorndike wrote that "the mind is a host of highly-particularized and independent faculties." Since the preceding objections seem to necessitate renouncing all claim that the tests measure general ability in the sense of anything that functions in a unitary manner, the natural step has been to take them as measuring some "average" or "sample" of many independent abilities. But here a crop of new difficulties has arisen. For it is objected that the operation of averaging, widespread and indispensable as it is both in psychology and elsewhere, still only becomes feasible when certain fundamental conditions are fulfilled. Certainly, any number of tests can be averaged. But can a person's abilities be so ? One of the con ditions evidently is that the domain should be settled within which the averaging is intended co run. No genuine averaging of a per son's abilities can be effected until there has first been some agreement as to whether memory, imagination, sensation, motor skill, etc., are to be counted in or not. No such agreement, as we have seen, appears to have been reached, or even to be pos sible. Another needful condition in order to obtain any genuine average is to settle on some definite unit. In the case of abilities no such unit would seem to be ever discoverable. A further con dition is that every unit should be given an equal weight, none being omitted and none counted more than once. This also can not ever be satisfied by any set of tests, it would seem ; by no method can they be so constructed that all the existing elements of ability are introduced and equally weighted. So much for averaging. If we turn to sampling, all the same difficulties recur, and indeed with some new ones added. For here again there is need of definite limits and definite units that are nowhere realiz able. Again, there is the apparently hopeless task of giving to every unit an equal weight, in the sense that the sampling should be so carried out as to afford every possible cognitive element an equal chance of being introduced into the tests.

Such, then, are the formidable difficulties against which the vogue of testing general ability has run its head. By these—in spite of its vast dimensions and its claims of _actual success—it has been compelled to pause and look around for some more secure foundation.

Re-interpretation.

All the time, however, that these ob stacles were becoming more and more stubbornly resistant to the efforts at relying solely on practical success in testing, other in vestigators more theoretically inclined were busily discovering means whereby such obstacles could be safely surmounted. In deed, the key to the solution had been supplied from the very beginning, namely, in taking general ability to be represented, not by any such psychological concept as "intelligence," but in stead by the statistical concept of g. That the latter is measurable was definitely demonstrated. Since it recurs throughout all the cognitive powers of each individual, the general ability was unequivocally designated as that—whatever it might be—which was measured by this individual constant.

But to put this standpoint upon a secure foundation, much further research was indispensable. One branch of progress was purely mathematical. The criterion of the divisibility into the two factors g and s had to be established more rigorously than was done at first. What was much more difficult, means had to be devised for counteracting the disturbances of this criterion by the so-called errors of sampling. The other great branch of ad vance besides the mathematical one was by way of broadening the scope of observation. Originally, the relevant work had been lim ited to a single experimenter dealing with little over a hundred in dividuals, all of whom were young boys. But already in 1912 the range had been extended to 14 experimenters treating 1,463 boys and girls, men and women, sane and insane. At the present day this number of experiment rs has certainly been quadrupled, whilst the number of individuals has been further multiplied at least twentyfold. All this progress, it must be added, has not been done without extensive and at times very acute controversy. This, however, has now mostly died down. The foregoing view is almost everywhere at any rate adopted in substance, if not always explicitly acknowledged.

So far we have only been considering the main theory, where the g is not credited with being anything more than a measured value which constantly recurs and therefore must have some constant basis in fact. By adopting this statistical standpoint, however, the psychological analysis is by no means rendered unnecessary; it is only, for convenience, postponed. Instead of having to be made prematurely in order to supply a definition to work with, it is accomplished gradually and provisionally in accordance with all such facts as may from time to time be elicited. The various psychological interpretations suggested for g constitute so many sub-theories of this. In respect of these, controversy not only still remains keen but probably will never be completely silenced. A fate may be expected for it analogous to that of the enquiry into the nature of electricity; for more than a century the latter has been susceptible of accurate measure ment, and yet to this day expert opinion as to its real nature is always changing and developing.

At the present moment, the interpretation of g which has been found to account best for all the known facts is to conceive it as measuring some general psychic or psycho-physical "energy" responsible for all the activities of which the mind is capable. The s's on this view become the various "engines" into any of which the energy may alternatively be directed, much as a current of electricity received from a sin gle power station may at one moment be directed to the turn ing of a wheel, at another to the raising of a lift, and then, say, to the heating of a furnace. This interpretation can be symbolized in the following diagram. The whole area represents the cerebral cortex; the shaded patch is any special group of neurons ; and the arrowheads stand for the lines of force from the entire cortex.

Among the most important of the engines has been found to be what is commonly called "mechanical aptitude," and what on psychological analysis is shown to consist of the power to educe spatial relations and correlates. Another engine (or system of engines) of cardinal importance is that which specifically sub serves linguistic ability. Further vital engines are those connected with the several sensory and motor abilities. Of the many facts discovered concerning all these engines, perhaps the most inter esting and significant is that they alone appear to be susceptible of improvement by practice; they alone, then, supply the material upon which education can work effectively. The amount of a person's "energy" would seem to be, under normal conditions, irrevocably fixed.

Still more important developments claimed for the theory of general and specific abilities have yet to be mentioned. For up to now we have only considered these in and by themselves. Whereas their investigation has led on to discovering some very impor tant further mental characteristics. In addition to the g which measures the quantity of the energy of the mind, there has also been determined a value p which measures the degree of the "inertia" of this energy, and, furthermore, an o which measures the degree that the supply of the energy tends to oscillate. These three measures conjointly are said to furnish the chief features distinguishing the abilities of one individual from those of another.

The evolution of the theory has gone further still, carrying it far beyond that whole sphere of psychology which appertains to individual differences. For during the time that the facts belong ing to this sphere have been investigated with the success de scribed, those belonging to "general" psychology have been treated from a fundamentally novel standpoint, that of the doctrine of "neogenesis." The unexpected result has been that these two spheres, hitherto kept by psychology in disastrous divorce from one another, would now seem to fit as hand and glove.

Finally, both these spheres of cognition—that of individual differences and that of general psychology—have in the continued development of the theory of g and s become organically united with that other great domain of the mind which is not cognitive but conative ; it does not deal with knowing, but with desiring and striving. To the g and the s's of the energy and the engines there has been added a w which brings all these into action, play ing thus the part of the "engineer." BIBLIOGRAPHY.-For a general account of the whole topic, see Bibliography.-For a general account of the whole topic, see The Abilities of Man (1927), by C. Spearman, in which will also be found a comprehensive bibliography ; also the work of A. Binet, Annee Psychologique (19o5) ; whilst the most notable contribution on the mathematical side has been that of Maxwell Garnett, Brit. Journ. Psych. (192o). Among other works which on no account should be overlooked are those of P. Ballard on the age at which g reaches its full growth, Brit. Journ. Psych. (1921) ; C. Burt on the develop ment of reasoning power, Journ. Educ. Psych. (1923) ; W. Stern in his Differential Psychology (191I) ; L. Thorndike on The Measure ment of Intelligence (1927) ; and of Webb on "Character and Intelli gence," Brit. Journ. Psych., Monthly Supplement (1915). (C. S.)