Chapter XV: Section I: Smell (10)

Having cited with some comments the extracts from Cudworth and Harris, Mr. James Mill observes, "Under the influence of such notions as these, men were led away from the real object of Classification, which remained, till a late period of metaphysical enquiry, not at all understood. Yet the truth appears by no means difficult to find, if we only observe the steps by which the mind acquires its knowledge, and the exigencies which give occasion to the contrivances to which it resorts" (p. 259).--He then proceeds, clearly and forcibly, to announce his own theory of classification, intended to dispel the mystery with which others have surrounded {275} it (p. 264). "The word _man_ is first applied to an individual: it is first associated with the idea of that individual, and acquires the power of calling up the idea of him: it is next applied to another individual, and acquires the power of calling up the idea of him: so of another and another, till it has acquired the power of calling up an indefinite number of those ideas indifferently. What happens? It does call up an indefinite number of the ideas of individuals, as often as it occurs: and calling them up in close combination, it forms them into a species of complex idea." "It thus appears that the word _man_ is not a word having a very simple idea, as was the opinion of the Realists: nor a word having no idea at all, as was that of the Nominalists: but a word calling up an indefinite number of ideas, by the irresistible laws of association, and forming them into one very complex and indistinct, but not therefore unintelligible, idea" (p. 265).--"As it is for the purpose of naming, and of naming with greater facility, that we form classes at all; so it is in furtherance of that same facility that such and such things only are included in one class, such and such things in another. Experience teaches us what sort of grouping answers this purpose best: under the suggestions of that experience, the application of a general word is tacitly and without much of reflection regulated: and by this process and no other, it is, that Classification is performed. It is the aggregation of an indefinite number of individuals, by their association with a particular name" (p. 268).--"It is Similarity or Resemblance, which, when we have applied a name to one individual, leads us to apply it to another and another till the whole forms an aggregate, connected together by the common relation of the aggregate to one and the same name" (p. 271).

Such is the theory of Mr. James Mill. Its great peculiarity is that it neither includes nor alludes to Abstraction. It admits in Classification nothing more than the one common name associated with an aggregate indefinite and indistinct, of similar concrete individuals. I shall now consider the manner {276} in which the Greek philosophers of the fourth century B.C. dealt with the same subject, and how far they merit the censure of having imported unnecessary mystery into it.

It is impossible to understand Plato unless we take our departure from his master Sokrates. Now it is precisely in regard to Classification, and the meaning and comprehension of general terms, that the originality and dialectical acuteness of Sokrates were most conspicuously manifested. He was the first philosopher (as Aristotle[c] tells us) who set before himself the Universal as an express object of investigation,--and who applied himself to find out and test the definition of universal terms. He wrote nothing; but he passed most part of his long life in public, and in talking indiscriminately with every one. Oral colloquy, and cross-examining interrogation, were carried by him to a pitch of excellence never equalled. Not only did he disclaim all power of teaching, but he explicitly avowed his own ignorance; professing to be a mere seeker of truth from others who knew better, and to be anxious only for answers such as would stand an accurate scrutiny. To this peculiar scheme--the topics on which he talked were adapted: for he avoided all recondite themes, and discussed only matters relating to man and society: such as What is the Holy? What is the Unholy? What are the Beautiful and the Mean the Just and Unjust? Temperance? Madness? Courage? Cowardice? A City? A man fit for citizenship? Command of Men? A man fit for commanding men? Such is the specimen-list given by Xenophon[d] of the themes chosen by Sokrates. We see that they are all general, and embodied in universal terms. But the terms as well as the themes were familiar to all: every man believed himself thoroughly to understand the meaning of the former--every one had convictions ready-made and decided on the latter. When Sokrates first opened the colloquy, respondents were surprised to be questioned about such subjects, upon which they presumed {277} that every one must know as well as themselves. But this confidence speedily vanished when they came to be tested by inductive[e] interrogatories: citation of appropriate particulars, included or not included in the generalities which they laid down. The result proved that they could not answer the questions without speedily contradicting themselves: that they did not understand the comprehension of their own universal terms: and that upon all these matters, on which they talked so confidently, they had never applied themselves deliberately to learn, nor could they say how their judgments had been acquired or certified.[f]

[Footnote c: Aristot. Metaphys. A. p. 987, b. 1, M. p. 1078, b. 30.]

[Footnote d: Xenophon, Memorab. I., 1--16.]

[Footnote e: So Aristotle calls them--[Greek: lo/gous e(paktikou/s].--Metaph. M. p. 1078, b. 28.]

[Footnote f: Xenophon, Memorab. IV. 2--13--30--36.]

The conviction formed in the mind of Sokrates, after long persistence in such colloquial cross-examination, is consigned in his defence before the Athenian judicature, pronounced a month before his death. He declared that what he found every where was real ignorance, combined with false persuasion of knowledge: that this was the chronic malady of the human mind, which it had been his mission to expose: that no man was willing to learn, because no man believed that he stood in need of learning: that, accordingly, the first step indispensable to all effective teaching, was to make the pupil a willing learner, by disabusing his mind of the false persuasion of knowledge, and by imparting to him the stimulus arising from a painful consciousness of ignorance.

Such was the remarkable psychological scrutiny instituted by Sokrates on his countrymen, and the verdict which it suggested to him. I have already observed that his great intellectual bent was to ascertain the definition of general terms, and to follow these out to a comprehensive and consistent classification.[g] It must be added that no man was ever less inclined to mysticism than Sokrates: and that he was thus {278} exempt from those misleading influences which (according to Mr. James Mill, p. 260) "have led men away from the real object of Classification, and prevented them from understanding it till a late period in metaphysical enquiry." Sokrates did not come before his countrymen with classifications of his own, originated or improved--nor did he teach them how the process ought to be conducted. His purpose was, to test and appreciate that Classification which he found ready-made and current among them. He pronounced it to be worthless and illusory.

[Footnote g: Xenophon. Memor. IV. 5, 12; IV. *6. 1--7--10--15. [Greek: ô(=n e(/neka skopô=n su\n toi=s sunou=sin, ti/ e(/kaston ei)/ê tô=n o)/ntôn, ou)de/pote e)/lêge].]

Now I wish to point out that what Sokrates thus depreciated, is exactly that which this Chapter of the Analysis lays before us as Classification generally. I agree with the Analysis that Classification, up to a certain point, grows out of the principle of Association and the exigencies of the human mind, by steps instructively set forth in that work. But such natural growth reaches no higher standard than that which Sokrates tested and found so lamentably deficient, even among a public of unusual intelligence. It does not deserve the name of a "mighty operation" (bestowed upon it by Mr. James Mill, p. 270). It is a rudimentary procedure, indispensable as a basis on which to build, and sufficing in the main for social communication, when no science or reasoned truth is required: but failing altogether to realise what has been understood by philosophers, from Sokrates downward, as the true and full purpose of Classification. So long as the Class is conceived to be only what the Analysis describes, an indistinct aggregate of resembling individuals denoted by the same name, without clearly understanding wherein the resemblance consists, or what facts and attributes are _connoted_ by the name[h] (I use the word _connote_, {279} not in the sense of the Analysis, but in the sense of Mr. John Stuart Mill)--so long will Classification continue to be, as Sokrates entitled it, a large persuasion of knowledge with little reality to sustain it.

[Footnote h: The necessity of determining the _connotation_ of the Class-term is distinctly put forward by Sokrates--Xenophon, Memorab. III. 14, 2. [Greek: lo/gô| o)/ntos peri\ o)noma/tôn, iph' oi(ô=| e)/rgô| e(/kaston ei)/ê--E)/choimen a)\n (e)\phê) ei)pei=n, e)pi\ poiô=| pote\ e)/rgô| a)/nthrôpos o)pso/phagos kalei=tai?] &c., also the remarkable passage IV., 6. 13--15, Plato, Sophistes, p. 218 B. [Greek: tou)/noma mo/non e)/chomen koinê=| to\ de\ e)/rgon, e)ph' ô(=| kalou=men], &c.]

I pass now from Sokrates to Plato. It is true, as we read in the Analysis, (p. 271) that Plato "was so deeply struck with the importance of Classification, that he seems to have regarded it as the sum of all philosophy." But what Plato thus admired was not the Classification that he found prevalent around him, such as this chapter of the Analysis depicts. Here Plato perfectly agreed with Sokrates. Among his immortal dialogues, several of the very best are devoted to the illustration of the Sokratic point of view: to the cross-examination and exposure of the minds around him, instructed as well as vulgar, in respect to the general terms familiarly used in speech. The Platonic questions and answers are framed to shew how little the respondents understand beneath those current generalities on which every one talks with confidence and fluency--and how little they can avoid contradiction or inconsistency, when their class-terms are confronted with particulars. In fact, Plato goes so far as to intimate that these uncertified classifications,-- generated in each man's mind by merely learning the application of words, and imbibed unconsciously, without special teaching, through the contagion of ordinary society--are rather worse than ignorance: inasmuch as they are accompanied by a false persuasion of knowledge. It would be (in the opinion of Plato) a comparative improvement, if this state of mental confusion, creating a false persuasion of knowledge, were broken up; and if there were substituted in place thereof positive ignorance, together with the naked and painful consciousness of being really ignorant. Only in this way could the mind of the learner be stimulated to active effort in the acquisition of genuine knowledge.[i]

[Footnote i: Plato, Sophistes, p. 230--231. Symposion, p. 204 A, Menon p. 84, A. D.]

Accordingly, when it is said that Plato was "deeply struck {280} with the importance of Classification," we must understand the phrase as applying to Classification, not as he found it prevalent, but as he idealized it. And the scheme that he imagined was not merely different from that which he found, but in direct repugnance to it. He denounced altogether the aggregate of individuals; he declared the class-constituent to reside in a reality apart from them, separate and self-existent--the Idea or Form. He enjoined the student of philosophy to fix his contemplation on these Class-Ideas, the real Realities, in their own luminous region: and for that purpose, to turn his back upon the phenomenal particulars, which were mere transitory, shadowy, incoherent projections of these Ideas[j]--and from the study of which no true knowledge could be obtained. Of the two statements in the Analysis--(p. 271) that "Plato never dreamed of the mystical visions of his successors," and that "his error (respecting Classification) lay in misconceiving the One; which he took, not for the aggregate, but something pervading the aggregate"--neither one nor the other appears to me accurate. In regard to the second of the two, indeed, you may find various passages of Plato which, if construed separately, would countenance it: for Plato does not always talk Realism--nor always consistently with himself. But still his capital and peculiar theory was, Realism. The Platonic One was not something pervading the aggregate of particulars, but an independent and immutable reality, apart from the aggregate: and Plato, when he thus conceived {281} the One, illustrating it by the vast hypotheses embodied in the Republic, Phædon, Phædrus, Symposion, Menon, &c., is the true originator of those "mystical visions" against which the Analysis justly protests. Such visions were doubtless suggested to Plato by "his deep sense of the importance of Classification:" but they are his own, though continued and amplified, without his decorative genius, by Neo-Platonic successors. His theory of classification was the first ever propounded; and that theory was Realism. The doctrine here ascribed to him by Mr. James Mill is much more Aristotelian than Platonic. The main issue raised by Aristotle against Plato was, upon the essential separation, and separate objective existence, of the Abstract and Universal: Plato affirmed it, Aristotle denied it.[k] Aristotle recognised no reality apart from the Particular, to which the Universal was attached as a predicate, either essential or accidental to its subject. The Aristotelian Universal may thus be called, in relation to a body of similar particulars, not the aggregate but something pervading the aggregate. But this is not Plato's view: it is the negation of the Platonic Realism.

[Footnote j: This is what we read in the memorable simile of the Cave, in Plato, Republic, VII., p. 514--519. The language used throughout this simile is [Greek: peria/gein, periakti/on, periagôgê/], &c. He supposes that the natural state of man is to have his face and vision towards the particular phenomena, and his back towards the universal realities: the great problem is, how to make the man face about, turn his back towards phenomena, and his eyes towards Universals--[Greek: ta\ o)/nta--ta\ noêta/]. Nothing can be learnt from observation however *acute, of the phenomena. The same point is enforced with all the charm of Platonic expression in Republ. V. 478, 479, VI., 493, 494. Symposion, p. 210--211, Phædon, p. 74--75.]

[Footnote k: According to Plato, it is [Greek: to\ e(\n _para\ ta\ polla/_]. According to Aristotle, it is [Greek: e(\n _kata\ pollô=n_--e(\n kai\ to\ au)to\ e)pi\ _pleio/nôn_ mê\ o(mô/numon e(\n _e)pi\_ pollô=n]. Analyt. Poster. I. 11, p. 77, a. 6. Metaphys. I. 9, p. 990, b. 7--13.

Whoever reads the portions of Plato's dialogues indicated in my last preceding foot note, will see how material this difference is between the two philosophers.

In the remarkable passage of the Analyt. Post. I. 24, p. 85, a. 30, b. 20, Aristotle notices the Platonic hypothesis that the Universal has real objective, separate, existence apart from its particulars ([Greek: to\ katho/lou e)pi\ ti para\ ta\ kath' e(/kasta]) as an illusion, mischievous and misleading--frequent, but not unavoidable.

See the antithesis between Plato and Aristotle, on the subject of Universals, more copiously explained in the recent work of Professor Bain, Mental and Moral Science, Appendix, pp. 6--20.]

When we read in the Analysis (p. 265) that "the word _man_ is not a word having a very simple idea, as was the opinion of the Realists; nor a word having no idea at all, as was that of the Nominalists" this language seems to me not well-chosen. {282} As to the Realists--the Platonic Ideas are conceived as eternal, immutable, grand, dignified, &c., but Aristotle[l] contends that they cannot all be simple: for the Idea of Man (e.g.) can hardly be simple, when there exists distinct Ideas of Animal and of Biped. As to the Nominalists--we cannot surely say that they conceived the universal term as "having no idea at all." A doctrine something like this is ascribed (on no certain testimony) to Stilpon, in the generation succeeding Aristotle: the word Man (Stilpon is said to have affirmed[m]) did not mean John more than William or Thomas or Richard, &c., therefore it did not mean either one of them: therefore it had no meaning at all. So also William of Ockham is said to have declared that Universal Terms were mere "flatus vocis:" but this (as Prantl has shewn[n]) was a phrase fastened upon him by his opponents, not employed by himself. Still less can it be admitted that Hobbes and Berkeley conceived the Universal Term as "having no idea at all." They denied indeed Universal Ideas in the Realistic sense: they also denied what Berkeley calls "determinate abstract Ideas:" but both of them explained (Berkeley especially) that the Universal term meant, any particular idea, considered as representing or standing for all other particular ideas of the same sort.[o] Whether this be the best and most complete explanation or not, it can hardly have been present to Mr. James Mill's mind, when he said that the Universal term had no idea at all in the opinion of the Nominalists.

[Footnote l: Aristot. Metaphys. Z. 1039, a. 27, 1040, a. 23.]

[Footnote m: See Grote, Plato and the other Companions of Sokrates, Vol. III., ch. 38, p. 523.]

[Footnote n: Prantl, Geschichte der Logik, Vol. III., Sect. 19, p. 327.]

[Footnote o: Berkeley, Principles of Human Knowledge, Introduction, Sect. 12, 15, 16.]

There is one other remark to be made, respecting the view of Classification presented in the eighth Chapter of the Analysis. We read in the beginning of that Chapter--p. 249--"Forming a class of things is a mode of regarding them. But what is meant by a mode of regarding things? This is mysterious: {283} and is as mysteriously explained, when it is said to be the taking into view the particulars in which individuals agree. For what is there which it is possible for the mind to take into view, in that in which individuals agree? Every colour is an individual colour, every size is an individual size, every shape is an individual shape. But things have no individual colour in common, no individual shape in common, no individual size in common: that is to say, they have neither shape, colour, nor size in common. What then is it which they have in common which the mind can take into view? Those who affirmed that it was something, could by no means tell. They substituted words for things: using vague and mystical phrases, which when examined meant nothing."

Here we find certain phrases, often used both in common speech and in philosophy, condemned us mystical and obscure. In the next or ninth Chapter (on Abstraction, p. 295 seq.), we shall see the language substituted for them, and the theory by which the mystery is supposed to be removed. I cannot but think that the theory of Mr. James Mill himself is open to quite as many objections as that which he impugns. He finds fault with those who affirm that the word _cube_ or _sphere_ is applied to a great many different objects by reason of the shape which they have in common; and that they may be regarded so far forth as _cube_ or _sphere_. But surely this would not have been considered as either incorrect or mysterious by any philosopher, from Aristotle downward. When I am told that it is incorrect, because the shape of each object is an _individual_ shape, I dissent from the reason given. In my judgment, the term _individual_ is a term applicable, properly and specially, to a concrete object--to that which Aristotle would have called a Hoc Aliquid. The term is not applicable to a quality or attribute. The same quality that belongs to one object, may also belong to an indefinite number of others. It is this common quality that is _connoted_ (in the sense of that word employed by Mr. John Stuart Mill) by the class-term: and if there were no common quality, the class-term would have no connotation. In other words, there would be no class: nor {284} would it be correct to apply to any two objects the same concrete appellative name.

But when we come to the following Chapter of the Analysis (ch. ix. on Abstraction, p. 296), we read as follows--"Let us suppose that we apply the adjective _black_ first to the word Man. We say 'black man.' But we speedily see that _for the same reason_ for which we say black man, we may say black horse, black cow, black coat, and so on. The word _black_ is thus associated with innumerable modifications of the sensation black. By frequent repetition, and the gradual strengthening of the association, these modifications are at last called up in such rapid succession that they appear commingled, and no longer many ideas, but one. _Black_ is therefore no longer an individual, but a general name. It marks not the particular black of a particular individual, but the black of every individual and of all individuals."

To say that we apply the word _black_ to the horse _for the same reason_ as we applied it to the man, is surely equivalent to saying that the colour of the horse is the same as that of the man: that blackness is the colour which they have in common. It is quite true that we begin by applying the name to one individual object, then apply it to another, and another, &c.; but always for the same reason--to designate (or _connote_, in the phraseology of Mr. John Stuart Mill) the same colour in them all, and to denote the objects considered under one and the same point of view. It may be that in fact there are differences in shade of colour: but the class-name leaves these out of sight. When we desire to call attention to them, we employ other words in addition to it. Every attribute is considered and named as One, which is or may be common to many individual objects: the objects only are individual.

It is to be regretted, I think, that Mr. James Mill disconnected Classification so pointedly from Abstraction, and insisted on explaining the former without taking account of the latter. Such disconnection is a novelty, as he himself states (p. 294): previous expositors thought that "abstraction was included in classification"--and, in my judgment, they were {285} right in thinking so, if (with Mr. James Mill) we are to consider Classification as a "great operation." An aggregate of concretes is not sufficient to constitute a Class, in any scientific sense, or as available in the march of reasoned truth. You must have, besides, the peculiar mode of regarding the aggregate: (a phrase which Mr. James Mill deprecates as mysterious, but which it is difficult to exchange for any other words more intelligible) you must have "that separating one or more of the ingredients of a complex idea from the rest, which has received the name of Abstraction"--to repeat the very just explanation given by him, p. 295--though that too, if we look at p. 249, he seems to consider as tainted with mystery.

We proceed afterwards to some clear and good additional remarks--p. 298. A class-term, as _black_, "is associated with two distinguishable things, but with the one much more than with the other: the clusters, with which it is associated, are variable: the peculiar sensation with which it is associated, is invariable. It is constantly, and therefore much more strongly, associated with the sensation, than with any of the clusters. It is at once a name of the clusters and a name of the sensation: but it is more peculiarly a name of the sensation." Again shortly afterwards, the abstract term is justly described as "marking exclusively one part (of the cluster), upon which such and such effects depend, no alteration being supposed in any other part of it."[p]

[Footnote p: The abstract term is coined for the express purpose of marking one part of a cluster simultaneously present to the mind, and fixing attention upon it without the other parts--but the concrete term is often made to serve the same purpose, by means of the adverb quatenus, [Greek: ka/thoson, ê(=|] &c. These phrases are frequent both in Plato and Aristotle: the stock of abstract terms was in their day comparatively small. It is needless to multiply illustrations of that which pervades the compositions of both: a very good one appears in Plato, Republ. I., p. 340 D, 341 C, 342.]

This process of marking exclusively, and attending to, one constant portion of a complex state of consciousness, amidst a {286} great variety of variable adjuncts--is doubtless one fundamental characteristic in Abstraction and Classification. A mystery was spread around it by Plato--first through his ascribing to the Constant a separate self-existence, apart from the Variables--still more by his hyperbolical predicates respecting these self-existent transcendental Entia. Plato[q] however in other passages gives many just opinions, respecting Classification, which are no way founded on Realism, and are equally admissible by Nominalists: and portions of Aristotle may be indicated, which describe the process of abstraction as clearly as any thing in Hobbes or Berkeley.[r]

[Footnote q: The two Platonic dialogues, Sophistes and Politikus, (in which processes of Classification are worked out,) give precepts, for correct and pertinent classification, not necessarily involving the theory of Realism, but rather putting it out of sight; though in one special part of the Sophistes, the debate is made to turn upon it. The main purpose of Plato is to fix upon some fact or phenomenon, clear and appropriate, as the groundwork for distinguishing each class or sub-class--and to define thereby each class-term (_i.e._, to determine its _connotation_, in the sense of Mr. John Stuart Mill). Plato deprecates the mere following out of resemblances as a most slippery proceeding ([Greek: o)listhêro/taton ge/nos]--Sophist. 231 A). The commonly received classes carry with them in his opinion, no real knowledge, but only the false persuasion of knowledge: he wants to break them up and remodel them.]

[Footnote r: See especially Aristot. De Memoriâ et Reminiscentiâ, c. 1, p. 449, b. 13. De Sensu et Sensili, c. 6, p. 445, b. 17. De Animâ III. 8, p. 432, a. 9.]

One farther remark may be made upon these two Chapters of the Analysis. Mr. James Mill seems to take little or no thought of Classification and Abstraction, except as performed by Adjectives. But the adjective presupposes a substantive, which is alike an appellative; and which has already performed its duty in the way of abstracting and classifying. This fact seems to be overlooked in the language of some sentences in the present Chapter: for example--"Some successions are found to depend upon the clusters called _objects_, all taken together. Thus a tree, a man, a stone, are the {287} antecedents of certain consequents, as such: and not on account of any particular part of the cluster. Other consequents depend not upon the whole cluster, but upon some particular part: thus a tall tree produces certain effects which a tree not tall cannot produce," &c.

I think that the phraseology of this passage is not quite clear. "The whole cluster all taken together" is not a tree as such--a man as such--a stone as such--but this particular man, tree, or stone, as it stands: John, Thomas, Caius or Titius, clothed with all his predicates, acting or suffering in some given manner. When we speak of a man _as such_ or _quatenus man_--we do not include the whole cluster, but only those attributes _connoted_ (in Mr. John Stuart Mill's sense of the word) by the name _man_: we speak of him as a member of the class _Man_. What I wish to point out is--That Man is a class-term, just as much as _tall_ or _short_: only it is the name of a larger class, while tall man is a smaller class under it. The school-logicians did not consider substantives as connotative, but only adjectives: Mr. James Mill has followed them as to this extent of the word, though he has inverted their meaning of it (see p. 299). Mr. John Stuart Mill, while declining to adopt the same inversion, has enlarged the meaning of the word _connotative_, so as to include appellative substantives as well as adjectives.--_G._]

[Editor's footnote 80: Rejecting the notion that classes and classification would not have existed but for the necessity of economizing names, we may say that objects are formed into classes on account of their resemblance. It is natural to think of like objects together; which is, indeed, one of the two fundamental laws of association. But the resembling objects which are spontaneously thought of together, are those which resemble each other obviously, in their superficial aspect. These are the only classes which we should form unpremeditatedly, and without the use of expedients. But there are other resemblances which are not superficially obvious; and many are not brought to light except by long experience, or observation carefully directed to the purpose; being mostly resemblances in the {288} manner in which the objects act on, or are acted on by, other things. These more recondite resemblances are often those which are of greatest importance to our interests. It is important to us that we should think of those things together, which agree in any particular that materially concerns us. For this purpose, besides the classes which form themselves in our minds spontaneously by the general law of association, we form other classes artificially, that is, we take pains to associate mentally together things which we wish to think of together, but which are not sufficiently associated by the spontaneous action of association by resemblance. The grand instrument we employ in forming these artificial associations, is general names. We give a common name to all the objects, we associate each of the objects with the name, and by their common association with the name they are knit together in close association with one another.

But in what manner does the name effect this purpose, of uniting into one complex class-idea all the objects which agree with one another in certain definite particulars? We effect this by associating the name in a peculiarly strong and close manner with those particulars. It is, of course, associated with the objects also; and the name seldom or never calls up the ideas of the class-characteristics unaccompanied by any other qualities of the objects. All our ideas are of individuals, or of numbers of individuals, and are clothed with more or fewer of the attributes which are peculiar to the individuals thought of. Still, a class-name stands in a very different relation to the definite resemblances which it is intended to mark, from that in which it stands to the various accessory circumstances which may form part of the image it calls up. There are certain attributes common to the entire class, which the class-name was either deliberately selected as a mark of, or, at all events, which guide us in the application of it. These attributes are the real meaning of the class-name--are what we intend to ascribe to an object when we call it by that name. With these the association of the name is close and strong: and the employment of the same name by different {289} persons, provided they employ it with a precise adherence to the meaning, ensures that they shall all include these attributes in the complex idea which they associate with the name. This is not the case with any of the other qualities of the individual objects, even if they happen to be common to all the objects, still less if they belong only to some of them. The class-name calls up, in every mind that hears or uses it, the idea of one or more individual objects, clothed more or less copiously with other qualities than those marked by the name; but these other qualities may, consistently with the purposes for which the class is formed and the name given, be different with different persons, and with the same person at different times. What images of individual horses the word horse shall call up, depends on such accidents as the person's taste in horses, the particular horses he may happen to possess, the descriptions he last read, or the casual peculiarities of the horses he recently saw. In general, therefore, no very strong or permanent association, and especially no association common to all who use the language, will be formed between the word horse and any of the qualities of horses but those expressly or tacitly recognised as the foundations of the class. The complex ideas thus formed consisting of an inner nucleus of definite elements always the same, imbedded in a generally much greater number of elements indefinitely variable, are our ideas of classes; the ideas connected with general names; what are called General Notions: which are neither real objective entities, as the Realists held, nor mere names, as supposed to be maintained by the Nominalists, nor abstract ideas excluding all properties not common to the class, such as Locke's famous Idea of a triangle that is neither equilateral nor isosceles nor scalene. We cannot represent to ourselves a triangle with no properties but those common to all triangles: but we may represent it to ourselves sometimes in one of those three forms, sometimes in another, being aware all the while that all of them are equally consistent with its being a triangle.

One important consequence of these considerations is, that {290} the meaning of a class-name is not the same thing with the complex idea associated with it. The complex idea associated with the name man, includes, in the mind of every one, innumerable simple ideas besides those which the name is intended to mark, and in the absence of which it would not be predicated. But this multitude of simple ideas which help to swell the complex idea are infinitely variable, and never exactly the same in any two persons, depending in each upon the amount of his knowledge, and the nature, variety, and recent date of his experience. They are therefore no part of the meaning of the name. They are not the association common to all, which it was intended to form, and which enables the name to be used by all in the same manner, to be understood in a common sense by all, and to serve, therefore, as a vehicle for the communication, between one and another, of the same thoughts. What does this, is the nucleus of more closely associated ideas, which is the constant element in the complex idea of the class, both in the same mind at different times, and in different minds.

It is proper to add, that the class-name is not solely a mark of the distinguishing class-attributes, it is a mark also of the objects. The name man does not merely signify the qualities of animal life, rationality, and the human form, it signifies all individual men. It even signifies these in a more direct way than it signifies the attributes, for it is predicated of the men, but not predicated of the attributes; just as the proper name of an individual man is predicated of him. We say, This is a man, just as we say, This is John Thompson: and if John Thompson is the name of one man, Man is, in the same manner, a name of all men. A class name, being thus a name of the various objects composing the class, signifies two distinct things, in two different modes of signification. It signifies the individual objects which are the class, and it signifies the common attributes which constitute the class. It is predicated only of the objects; but when predicated, it conveys the information that these objects possess those attributes. Every concrete class-name is thus a connotative name. It marks {291} both the objects and their common attributes, or rather, that portion of their common attributes in virtue of which they have been made into a class. It _denotes_ the objects, and, in a mode of speech lately revived from the old logicians, it _connotes_ the attributes. The author of the Analysis employs the word connote in a different manner; we shall presently examine which of the two is best.

We are now ready to consider whether the author's account of the ideas connected with General Names is a true and sufficient one. It is best expressed in his own words. "The word Man, we shall say, is first applied to an individual; it is first associated with the idea of that individual, and acquires the power of calling up the idea of him; it is next applied to another individual, and acquires the power of calling up the idea of him; so of another, and another, till it has become associated with an indefinite number, and has acquired the power of calling up an indefinite number of those ideas indifferently. What happens? It does call up an indefinite number of the ideas of individuals, as often as it occurs, and calling them up in close connexion, it forms them into a species of complex idea. . . . When the word man calls up the ideas of an indefinite number of individuals, not only of all those to whom I have individually given the name, but of all those to whom I have in imagination given it, or imagine it will ever be given, and forms all those ideas into one,--it is evidently a very complex idea, and therefore indistinct; and this indistinctness has doubtless been the main cause of the mystery which has appeared to belong to it. That this however is the process, is an inevitable result of the laws of association."

In brief, my idea of a Man is a complex idea compounded of the ideas of all the men I have ever known and of all those I have ever imagined, knit together into a kind of unit by a close association.

The author's description of the manner in which the class-association begins to be formed, is true and instructive; but does any one's idea of a man actually include all that the author {292} finds in it? By an inevitable result of the laws of association, it is impossible to form an idea of a man in the abstract; the class-attributes are always represented in the mind as part of an image of an individual, either remembered or imagined; this individual may vary from time to time, and several images of individuals may present themselves either alternatively or in succession: but is it necessary that the name should recal images of all the men I ever knew or imagined, or even all of whom I retain a remembrance? In no person who has seen or known many men, can this be the case. Apart from the ideas of the common attributes, the other ideas whether of attributes or of individual men, which enter into the complex idea, are indefinitely variable not only in kind but in quantity. Some people's complex idea of the class is extremely meagre, that of others very ample. Sometimes we know a class only from its definition, i.e. from an enumeration of its class-attributes, as in the case of an object which we have only read of in scientific books: in such a case the idea raised by the class-name will not be limited to the class-attributes, for we are unable to conceive any object otherwise than clothed with miscellaneous attributes: but these, not being derived from experience of the objects, may be such as the objects never had, nor could have; while nevertheless the class, and the class-name, answer their proper purpose; they cause us to group together all the things possessing the class-attributes, and they inform us that we may expect those attributes in anything of which that name is predicated.

The defect, as it seems to me, of the view taken of General Names in the text, is that it ignores this distinction between the meaning of a general name, and the remainder of the idea which the general name calls up. That remainder is uncertain, variable, scanty in some cases, copious in others, and connected with the name by a very slight tie of association, continually overcome by counter-associations. The only part of the complex idea that is permanent in the same mind, or common to several minds, consists of the distinctive attributes marked by the class-name. Nothing else is universally present, though {293} something else is always present: but whatever else be present, it is through these only that the class-name does its work, and effects the end of its existence. We need not therefore be surprised that these attributes, being all that is of importance in the complex idea, should for a long time have been supposed to be all that is contained in it. The truest doctrine which can be laid down on the subject seems to be this--that the idea corresponding to a class-name is the idea of a certain constant combination of class attributes, accompanied by a miscellaneous and indefinitely variable collection of ideas of individual objects belonging to the class.--_Ed._]

{294} CHAPTER IX.

ABSTRACTION.

"I think, too, that he (Mr. Locke) would have seen the advantage of 'thoroughly weighing,' not only (as he says) 'the imperfections of Language;' but its _perfections_ also: For the perfections of Language, not properly understood, have been one of the chief causes of the imperfections of our knowledge."--_Diversions of Purley_, _by John Horne Tooke, A.M._, i. 37.

THE two cases of Consciousness, CLASSIFICATION, and ABSTRACTION, have not, generally, been well distinguished.

According to the common accounts of Classification, ABSTRACTION was included in it. When it is said, that, in order to classify, we leave out of view all the circumstances in which individuals differ, and retain only those in which they agree; this separating one portion of what is contained in a complex idea, and making it an object of consideration by itself, is the process which is named Abstraction, at least a main part of that process.

It is necessary now to inquire what are the purposes to which this separating of the parts of a complex idea, and considering and naming the separated parts by themselves, is subservient.

{295} We have already observed the following remarkable things in the process of naming: 1, Assigning names of those clusters of ideas called objects; as man, fish; 2, Generalizing those names, so as to make them represent a class; 3, Framing adjectives by which minor classes are cut out of larger.

Those adjectives are all names, of some separate portion of a cluster, and are, therefore, all instruments of abstraction, or of that separating one or more of the ingredients of a complex idea from the rest, which has received the name of Abstraction. One purpose of Abstraction, therefore, is the formation of those _sub-species_, the formation of which is required for certain purposes of speech.

These observations will be rendered familiar by examples. We say, tall man, red flower, race horse. In my complex idea of a man, or the cluster of ideas of sense to which I affix that mark, are included, certain ideas of colour, of figure, size, and so on. By the word tall, I single out a portion of those ideas, namely, the part relating to size, or rather size in one direction, and mark the separation by the sign or name. In my complex idea of a flower, colour is always one of the ingredients. By applying the adjective red, I single out this one from the rest, and point it out for peculiar consideration. The explanation is obvious, and need not be pursued in a greater number of instances.

Words of this description all denote differences; either such as mark out species from genera, or such as mark out individuals from species. Of this latter sort the number is very small; of which the reason is obvious; individual differences are too numerous to {296} receive names, and are marked by contrivances of abridgment which will be spoken of hereafter.

To explain this notation of differences; the same examples will suffice. In the phrase "tall man," the adjective "tall" marks the difference between such a man, and "short man," or "middle-sized man." Of the genus man, tall men are one species; and the difference between them and the rest of the genus is marked by the word tall. Of the genus flower, red flowers form a species, and the difference between them and the rest of the genus is marked by the adjective red. Of the genus horse, race horse forms a species, and the difference between this species and the rest of the genus is marked by the word race.

It is of importance further to observe, that adjectives singling out ideas which are not differences, that is, ideas common to the whole class, are useless: as, tangible wood; coloured man; sentient animal. Such epithets express no more than what is expressed by the name without them.

Another thing requiring the attention of the student is the mode in which these differential adjectives are generalized. As the word man, applied first to one individual, then to another, becomes associated with every individual, and every variety of the species, and calls them all up in one very complex idea; so are these adjectives applied to one class after another, and by that means at last call up a very complicated idea. Let us take the word "black" for an example; and let us suppose that we apply this adjective first to the word man. We say "black man." But we speedily see that for the same reason for which we say black man we may say black horse, black cow, {297} black coat, and so on. The word black is thus associated with innumerable modifications of the sensation black. By frequent repetition, and the gradual strengthening of the association, these modifications are at last called up in such rapid succession that they appear commingled, and no longer many ideas, but one. Black is therefore no longer an individual but a general name. It marks not the particular black of a particular individual; but the black of every individual, and of all individuals.[81] The same is the case {298} with all other words of the same class. Thus I apply the word sweet, first to the lump of sugar in my mouth, next to honey, next to grapes, and so on. It thus becomes associated with numerous modifications of the sensation sweet; and when the association is sufficiently strengthened by repetition, calls them up in such close succession, that they are converted into one complex idea. We are also to remember, that the idea and the name have a mutual power over one another. As the word black calls up the complex idea, so every modification of black calls up the name; and in this, as in other cases, the name actually forms a part of the complex idea.

[Editor's footnote 81: The example which the author has here selected of a general name, sets in a strong light the imperfection of the theory of general names, laid down by him in the preceding chapter. A name like "black," which marks a simple sensation, is an extreme case of the inapplicability of the theory. Can it be maintained that the idea called up in our minds by the word black, is an idea compounded of ideas of black men, black horses, black cows, black coats, and the like? If I can trust my own consciousness, the word need not, and generally does not, call up any idea but that of a single black surface. It is still not an abstract idea, but the idea of an individual object. It is not a mere idea of colour; it is that, combined with ideas of extension and figure, always present but extremely vague, because varying, even from one moment to the next. These vague ideas of an uncertain extension and figure, combined with the perfectly definite idea of a single sensation of colour, are, to my consciousness, the sole components of the complex idea associated with the word black. I am unable to find in that complex idea the ideas of black men, horses, or other definite things, though such ideas may of course be recalled by it.

In such a case as this, the idea of a black colour fills by itself the place of the inner nucleus of ideas knit together by a closer association, which I have described as forming the permanent part of our ideas of classes of objects, and the meaning of the class-names.--_Ed._]

The next thing, which I shall observe, deserves in a high degree, the attention of the learner. In the various applications of that species of marks which we are now considering, they are associated with two distinguishable things; but with the one much more than the other. Thus, when we say black man, black horse, black coat, and so of all other black things, the word black is associated with the cluster, man, as often as black man is the expression; with the cluster horse, as often as black horse is the expression, and so on with infinite variety: but at the same time that it is associated with each of those various clusters, it is also associated with the peculiar sensation of colour which it is intended to mark. The CLUSTERS, therefore, with which it is associated, are variable; the PECULIAR SENSATION with which it is associated is invariable. It is much more constantly, and therefore much more strongly associated with the SENSATION than with any of the CLUSTERS. It is at once a name of the clusters, and a name of the {299} sensation; but it is more peculiarly a name of the SENSATION.

We have, in a preceding note, observed, that such words have been called _connotative_; and I shall find much convenience in using the term NOTATION to point out the sensation or sensations which are peculiarly marked by such words, the term CONNOTATION to point out the clusters which they mark along with this their principal meaning.

Thus the word, black, NOTES that of which black is more peculiarly the name, a particular colour; it CONNOTES the clusters with the names of which it is joined: in the expression, black man, it connotes man; black horse, it connotes horse; and so of all other cases. The ancient Logicians used these terms, in the inverse order; very absurdly, in my opinion.[82]

[Editor's footnote 82: The word Connote, with its substantive Connotation, was used by the old logicians in two senses; a wider, and a narrower sense. The wider is that in which, up to this place, the author of the Analysis has almost invariably used it; and is the sense in which he defined it, in a note to section *5 of his first chapter. "There is a large class of words which denote two things both together; but the one perfectly distinguishable from the other. Of these two things, also, it is observable, that such words express the one primarily as it were; the other in a way which may be called secondary. Thus white, in the phrase white horse, denotes two things, the colour and the horse; but it denotes the colour primarily, the horse secondarily. We shall find it very convenient to say, therefore, that it _notes_ the primary, _connotes_ the secondary signification."

This use of terms is attended with the difficulty, that it may often be disputed which of the significations is primary and which secondary. In the example given, most people would agree with the author that the colour is the primary signification; the word being associated with the objects, only through its previous association with the colour. But take the other of the two words, horse. That too is connotative, and in the same manner. It signifies any and every individual horse, and it also signifies those attributes common to horses, which led to their being classed together and receiving that common name. Which, in this case, is the primary, and which the secondary signification? The author would probably say, that in this case, unlike the other, horse is the primary signification, the attributes the secondary. Yet in this equally with the former case, the attributes are the foundation of the meaning: a thing is called a horse to express its resemblance to other horses; and the resemblance consists of the common attributes. The question might be discussed, pro and con, by many arguments, without any conclusive result. The difference between primary and secondary acceptations is too uncertain, and at best too superficial, to be adopted as the logical foundation of the distinction between the two modes of signification.

The author, however, has, throughout the preceding chapters, regarded words as _connoting_ any number of things which though included in their signification, are not, in his judgment, what they primarily signify. He said, for example, that a verb notes an action, and connotes the agent (as either me, thee, or some third person), the number of agents (as one or more), the time (as past, present, or future), and three modes, "that in which there is no reference to anything preceding, that in which there is a reference to something preceding, and that in which reference is made to the will of one of the Persons." I cite this complicated case, to shew by a striking example the great latitude with which the author uses the word Connote.