The Online Library of Liberty

A project of Liberty Fund, Inc.

• Introduction
• Textual Introduction
• Preface to the Third Edition[*]
• Chapter I: Introductory Remarks
• Chapter II: The Relativity of Human Knowledge
• Chapter III: The Doctrine of the Relativity of Human Knowledge, As Held By Sir William Hamilton
• Chapter IV: In What Respect Sir William Hamilton Really Differs From the Philosophers of the Absolute
• Chapter V: What Is Rejected As Knowledge By Sir William Hamilton, Brought Back Under the Name of Belief
• Chapter VI: The Philosophy of the Conditioned
• Chapter VII: The Philosophy of the Conditioned, As Applied By Mr. Mansel to the Limits of Religious Thought
• Chapter VIII: Of Consciousness, As Understood By Sir William Hamilton
• Chapter IX: Of the Interpretation of Consciousness
• Chapter X: Sir William Hamilton’s View of the Different Theories Respecting the Belief In an External World
• Chapter XI: The Psychological Theory of the Belief In an External World
• Chapter XII: The Psychological Theory of the Belief In Matter, How Far Applicable to Mind
• Chapter XIII: The Psychological Theory of the Primary Qualities of Matter
• Chapter XIV: How Sir William Hamilton and Mr. Mansel Dispose of the Law of Inseparable Association
• Chapter XV: Sir William Hamilton’s Doctrine of Unconscious Mental Modifications
• Chapter XVI: Sir William Hamilton’s Theory of Causation
• Chapter XVII: The Doctrine of Concepts, Or General Notions
• Chapter XVIII: Of Judgment
• Chapter XIX: Of Reasoning
• Chapter XX: On Sir William Hamilton’s Conception of Logic As a Science. Is Logic the Science of the Laws, Or Forms, of Thought?
• Chapter XXI: The Fundamental Laws of Thought According to Sir William Hamilton
• Chapter XXII: Of Sir William Hamilton’s Supposed Improvements In Formal Logic
• Chapter XXIII: Of Some Minor Peculiarities of Doctrine In Sir William Hamilton’s View of Formal Logic
• Chapter XXIV: Of Some Natural Prejudices Countenanced By Sir William Hamilton, and Some Fallacies Which He Considers Insoluble
• Chapter XXV: Sir William Hamilton’s Theory of Pleasure and Pain
• Chapter XXVI: On the Freedom of the Will
• Chapter XXVII: Sir William Hamilton’s Opinions On the Study of Mathematics
• Chapter XXVIII: Concluding Remarks
• Appendices
• Appendix A: Manuscript Fragments
• Appendix B: Textual Emendations
• Appendix C: Corrected References
• Appendix D: Bibliographic Index of Persons and Works Cited In the Examination, With Variants and Notes

Textual Introduction

JOHN M. ROBSON

I.

BACKGROUND

an examination of sir william hamilton’s philosophy is in several respects exceptional among Mill’s works. Although he devoted several major essays (such as “Bentham” and “Coleridge”), and one book (Auguste Comte and Positivism—originally a pair of essays) to individuals, only here did he subject an author’s texts to a searching and detailed analysis, sustained by an admitted polemical intent. Only part of the work is devoted to an exposition of Mill’s own views, and a few passages at most could be said to provide the kind of synthesis so typical of his other major writings. The kinds of revisions revealed by collation of the editions are also unusual in two related respects: a much higher proportion than in his other works is devoted to answering critics; and far more of the changes are in the form of added footnotes than is usual for him. Another difference is that the response to the book was immediate and strong: it elicited more reviews and critical replies in a short period of time than his Principles of Political Economy, System of Logic, and even On Liberty. Published in 1865, the first edition (of 1000 copies) sold out so quickly that a second edition was prepared within a couple of months, and a third edition, which was published two years after the first, would have appeared sooner had Mill not wished to answer his critics fully and at leisure. A fourth edition, the last in his lifetime, appeared in 1872 only five years after the third, and the work continued in demand for about twenty years.¹

As will be shown below, the evidently controversial nature of the argument explains much of the demand for the Examination; to some extent, however, Mill himself became more widely known at this time. His election campaign of 1865, though it came after both the first and second editions, must have increased the sales to troubled opponents as well as supporters. Also, the extraordinary interest in his other writings in these years added to, as well as reflected, his new prominence.²

The content and form of the argument is best seen against at least a brief outline of Mill’s interest in and acquaintance with Hamilton’s writings—they did not meet one another or, evidently, correspond. Sir William Hamilton (1788-1856), like Mill, was widely known long before any writings appeared under his name; indeed, unlike Mill, he began publishing significant articles anonymously only in his early forties. Described in 1814 to De Quincey by John Wilson as a “monster of erudition,”³ and remembered as a student at Oxford for his unexampled knowledge of obscure commentators on the Classics, he was elected Professor of Civil History at Edinburgh in 1821, but can hardly have become famous in that capacity, as the emolument soon ceased and he stopped lecturing. In 1829 appeared the first of his fifteen articles in the Edinburgh Review, his review of Cousin; one can probably assume that the tribal telegraph began to send the message that Hamilton was “coming out,” and Mill in London may soon have known; the Cousin review, coincidentally, appeared in the same number (Vol. L, October, 1829) as the third of Macaulay’s attacks on James Mill and Utilitarianism, and so it is almost certain that the younger Mill saw it, even if he did not know who had written it. In any case, the earliest extant reference comes in a letter from Mill to Carlyle of 2 August, 1833, in which he mistakenly assumes that Sir William Hamilton is the “strangest old schoolman (in a new body only forty years old)” to whom Carlyle had talked in the preceding winter. Mill’s assumption may have been founded on knowledge that Hamilton was the author of the erudite (but undoubtedly not to Mill persuasive) “Recent Publications on Logical Science” in the Edinburgh for April, 1833. Carlyle corrected Mill, saying that he had meant “a ganz ausgestorbener Mann,” considerably inferior to Hamilton, whom he also had met.⁴ It seems very likely that in the next year, after moving to London, Carlyle is referring to the proposed London Review and to Mill when he writes to Hamilton to say that there is talk of founding “a new periodical, gn another than the bibliopolic principle, with intent to show Liberalism under a better than its present rather sooty and ginshop aspect,” and that having been asked whether Hamilton might write for it, had “answered, Possible.” Hamilton, a strong Whig, writing later to Sarah Austin, indicates cautious interest in such a connection, but says his help could at best be occasional: “. . . I am too much occupied with matters apart from all popular interest, and have in the ‘Edinburgh Review’ an outlet more than sufficient for any superfluous energy with which I may be distressed.”⁵ In the event, Hamilton did not contribute to the London Review (or the London and Westminster), but one may assume that Mill was aware of him from this time on, and would know of his widely discussed election to the Chair of Logic and Metaphysics at Edinburgh in 1836. Mill, however, makes no direct allusion to Hamilton until 1842, when, having virtually finished his Logic, he speculates that, if John Austin does not review it for the Edinburgh, it is likely that Hamilton will, in a manner “hostile, but intelligent.”⁶ Still Hamilton had not published a book, but in 1846, ten years later than he had anticipated, his edition of Thomas Reid’s Works appeared, packed with his own footnotes and supplementary dissertations (the latter oddly and confusingly incomplete, as we shall see). Though he had suffered a severe stroke in 1844, he continued to lecture, and in 1852 published a collection of his review articles, Discussions on Philosophy and Literature, Education and University Reform. Mill, who owned the second edition (1853), obviously read it soon after its publication, for he added references to it in the fourth edition of his Logic (1856, the year of Hamilton’s death). In 1859 Mansel’s two-volume edition of Hamilton’s Lectures on Metaphysics appeared, followed in 1860 by the companion two-volume Lectures on Logic.⁷

II.

THE WRITING OF THE EXAMINATION

at this point, one may cite Mill’s account in his Autobiography of his reasons for turning to Hamilton’s philosophy as a subject. (This account, it should be noted, was written in 1869-70, that is, in the years between the third and fourth editions of his Examination.) He was at that time seeking a subject, feeling, apparently, that he had completed, at least for the time being, all he was able to do of the writing programme he and Harriet had agreed on in the 1850s.⁸ In particular, Considerations on Representative Government (first and second editions) and Utilitarianism (in its periodical form) had appeared in 1861, and The Subjection of Women, presumably in almost its final form, had been put aside in readiness for a more propitious occasion for publication. He wrote in Avignon in January, 1862, “The Contest in America,” and, after a seven-month trip to Greece and Turkey, in September (one must assume) composed, back in Avignon, a review of Cairnes’s The Slave Power.⁹

In the Autobiography, after mentioning the latter article, he says that the Examination was his “chief product” during the “next two years.” He had, however, begun serious study and consideration of Hamilton a year earlier, when he read Hamilton’s Lectures (which he erroneously dates as 1860 and 1861) “towards the end of the latter year, with a half formed intention of giving an account of them in a Review”;¹⁰ in fact, he wrote to Alexander Bain in November, 1861, saying that he intended to “take up Sir William Hamilton,” and try to make an article on him for the Westminster Review.¹¹ However, he soon decided (actually, within about a month)¹² that to do so “would be idle,” for “justice could not be done to the subject in less than a volume.” But should he write such a work? On reflection, he thought he should. As he indicates, up to this time he “had not neglected” the Discussions in Philosophy,¹³ though he had postponed study of the “Notes to Reid” because of “their unfinished state. . . .” Actually, it was not the “Notes” (Hamilton’s erudite and lengthy footnotes to passages in his edition of Reid), but the “Supplementary Dissertations” added at the end of the volume that were incomplete.

The story is a very pecular and confusing one: for reasons that are inadequately given by Mansel in the sixth edition or by Veitch in the Memoir of Sir William Hamilton, the first edition, prepared by Hamilton himself, breaks off (as Mill indicates, 3 below) in the middle of a sentence in Note D*** of the “Dissertations,” on 914; stereotyped editions appeared in the same form until 1863 (seven years after Hamilton’s death), when a sixth (also stereotyped) edition, prepared by Mansel, had a completion (after an insertion in square brackets) of Note D***, and further material.¹⁴ Though intriguing as a bibliographic puzzle, this curiosity would not be worth dwelling on here, had not Mill contributed to the difficulty by mentioning (33n), in a passage added in his fourth edition (1872), that his attention had been drawn to a section (Note N, itself “unfortunately left unfinished”) in “the posthumous continuation” of the “Dissertations,” and so suggesting, in conjunction with his earlier remark that the work was incomplete, that he had not seen the sixth (expanded) edition of Reid’s Works until he was preparing his own fourth edition. And Mill added in 1872 another note (255n) quoting from the additional material, again hinting that he had just come across it. However, Mill in fact was aware of the sixth edition when he wrote the Examination, for he quotes from the added material in his first edition (1865), mentioning that the passage comes from “one of the fragments recently [i.e., in 1863] published by his editors, in continuation of the Dissertations on Reid” (117). And, referring to Note D***, he says in the first edition, “this Dissertation . . . originally broke off abruptly, but the conclusion . . . has recently been supplied from the author’s papers . . .” (251n). Indeed, the first reference in the Examination includes the observation that the “Dissertations” leave off, “scarcely half finished,” in midsentence; to make this judgment, he must have had the other “half” before his eyes.

In any case, it would seem likely that Mill did not carefully study Hamilton’s edition of Reid until after his reading of the Lectures in late 1861. That reading was to him disappointing, for the Discussions, containing Hamilton’s “vigorous polemic against the later Transcendentalists, and his strenuous assertion of some important principles, especially the Relativity of human knowledge,” had attracted Mill’s sympathy and admiration, much as he realized the difference between himself and Hamilton concerning the bases of mental philosophy.¹⁵ “His Lectures,” says Mill,

Mill goes on, in a passage of intensity and force, to explain why Hamilton, a man of “imposing character” and “great personal merits and mental endowments,” came to embody for him his most resolute enemies, the Intuitionists (he makes special reference to Mansel, paraphrasing his attack in the Examination on the immorality of Mansel’s view of God), and so to justify “a thorough examination of all [Hamilton’s] most important doctrines, and an estimate of his general claims to eminence as a philosopher.” Or, in stronger language: “there ought to be a hand-to-hand fight between [the school of Intuition and the school of Experience and Association], . . . controversial as well as expository writings were needed, and . . . the time was come when such controversy would be useful.”¹⁷ As he had said to Bain in December, 1861, after having “studied all Sir W. Hamilton’s works pretty thoroughly”: “The great recommendation of this project is, that it will enable me to supply what was prudently left deficient in the Logic, and to do the kind of service which I am capable of to rational psychology, namely, to its Polemik.”¹⁸ Much the same attitude was conveyed to George Grote on 10 January, 1862:

Presumably he put aside Hamilton during the long trip to Greece referred to above, but with characteristic energy and thoroughness he was back at the task before the end of the year, mentioning in December to Theodor Gomperz his interest in Gomperz’s work on the principle of contradiction, for he had “commenced writing something to which a full understanding of that subject is indispensable,” and he had not yet thoroughly mastered it.²⁰ Bain says (without specific dates) that Mill, who was regularly corresponding with him at the time, “read all Hamilton’s writings three times over; and all the books that he thought in any way related to the subjects treated of.”²¹ These included, by early 1863, Mansel’s Limits of Religious Thought (a “detestable, . . . absolutely loathsome book”) and (re-read) Ferrier’s Institutes.²² The year of 1863 was not busy by Mill’s standards, his only major article being “Austin on Jurisprudence” in the October Edinburgh, and the only edition being the first book version of Utilitarianism. He spent April and May in Avignon, and then spent the next months in London (with a few days botanizing); he was busy enough socially in those months to express relief to Gomperz on 5 July that his life was “about to relapse into its usual wholesome tranquillity,” adding: “. . . I have been enabled to have a few days work at my book on Hamilton with which I now mean to persevere steadily.”²³ Returning to Avignon in early September, he was able to tell John Chapman on 5 October that, having finished his review of Austin, he was “at present chiefly writing on metaphysics.”²⁴ To Bain he said on 22 November that he had finished the book, “as far as regards the first writing,” and would not start rewriting until he had seen Bain’s more “matured form” of the analysis of primary qualities (i.e., in the second edition of his The Senses and the Intellect).²⁵ And again, on 4 December, he reports to Henry Fawcett: “. . . I have had little time to think on any scientific subject except Metaphysics, on which I am making good progress in the work I am about.”²⁶

It is probably to the work of this period that Mill and Bain refer as occasioning Mill’s decreased respect for Hamilton after the careful study of his writings. Mill says: “As I advanced in my task, the damage to Sir W. Hamilton’s reputation became greater than I at first expected, through the almost incredible multitude of inconsistencies which shewed themselves on comparing different passages with one another.”²⁷ Bain’s version is similar: “His picture of Hamilton grew darker as he went on; chiefly from the increasing sense of his inconsistencies. He often wished that Hamilton were alive to answer for himself.”²⁸ This coincidence is not surprising, of course, for Bain had the Autobiography by him, as well as Mill’s letter of 22 November, 1863, in which the tone is even sharper:

In fact this letter gives us the best picture of Mill’s progress. Enclosing a table of contents (now lost), he says that on all these heads he has “written chapters which are not unfit to print even now,” though he is, on the basis of “a third consecutive reading of Hamilton’s philosophical writings from beginning to end,” making “notes for additions & improvements” on the “blank pages” (i.e., the versos) of the manuscript. And he continued with his reading, asking Bain for information about Immanuel von Fichte, Vogt, and Moleschott (none of whom, incidentally, was demonstrably to influence his views).³⁰

The next year, 1864, also saw little publication by Mill, with no major essays or new works, only the second edition of Utilitarianism and the third of On Liberty appearing (both with the most trivial of revisions), and it may reasonably be argued that most of his working time in the first half of the year was given to rewriting the Examination, both in Avignon and London.³¹ Writing to Bain on 18 March, 1864, to thank him for the second edition of his The Senses and the Intellect, Mill says that the “remaining portion” of the Examination will—presumably as a result of Bain’s work—“now be plain sailing.” And, after discussing related matters at length, he concludes by saying that he hopes to have “at least some chapters of the Hamilton in a state to shew” to Bain in June.³² He notified Gomperz in June that, after hard work, the book was “well advanced towards completion,”³³ and he was able to let Bain read “the finished MS. of a large part of the book,” on which Bain made “a variety of minor suggestions,” and Mill “completed the work for the press the same autumn.”³⁴ Though we do not know when he approached Longman, by late October he told Augustus De Morgan that he anticipated publication in the spring of 1865,³⁵ and his attention had turned to his articles on Comte, which were finished in February, and appeared in the Westminster for April and July, 1865.

The Examination was published in an edition of 1000 copies on 13 April,³⁶ and by the end of the month had sold four hundred copies;³⁷ a second edition also of 1000 copies was called for, revised, printed, and published by 24 July.³⁸ As the surviving correspondence and the printed record demonstrate, Mill soon was engaged in replying to friend and foe, and the debate, private and public, continued for some years. He wrote, during its later phase:

The year 1865 having been extremely busy for Mill, 1866 was even more demanding, as his parliamentary duties, which for him meant constant mental as well as physical presence, speeches, and heavy responsibilities outside the House in connection with the Jamaica Committee and the Hyde Park riots, occupied a great deal of his time. In that year also his “Grote’s Plato,” a short book in itself, appeared, as did the slightly revised second edition of Auguste Comte and Positivism. But he found time to read and consider the responses to the Examination, and to report on them to Grote⁴⁰ and to Bain, the latter of whom says that Mill, after the close of the session in August, and a subsequent tour of the Alps and Pyrenees, settled down in Avignon to write his Rectorial Address for St. Andrews, and “to answer the attacks on Hamilton for the third edition; both which feats he accomplished before the opening of the session of 1867”⁴¹ in February. Mill was aware of the need for a third edition in April of 1866,⁴² but (with Longman’s concurrence) decided not to rush the rewriting,⁴³ and had “got through fully three fourths of the revision” by the end of the year.⁴⁴ Though the edition (again of 1000 copies) was not published until May,⁴⁵ it seems likely that, as Bain says, he had finished the revision before his return to London for the session, because early in February he told W. G. Ward, towards whom he always showed more than courtesy, that he would not be able in the revision to take account of Ward’s “Science, Prayer, Free Will, and Miracles,” even if he immediately saw proof of it.⁴⁶

The volume continued to sell, though more slowly: as Longman Division Books show, by June, 231 copies were disposed of, and in the next twelve months, till June, 1868, another 232. In the following twelve-month periods 162, 161, 141, and 148 were sold, so that by June, 1872 (what with some wastage and copies otherwise distributed), there were only twenty-seven copies left.

Further replies and discussions appeared in these years, and the French translation by Cazelles, published in 1869, brought forth notices in France. Mill proceeded with the substantial task of replying to critics, presumably reading and pondering the responses as they appeared. When he turned his hand to the actual revision we do not know, it being likely that, as usual, he waited until it was evident that a new edition was needed, which, as the account books suggest, was probably during 1871, there being only 176 copies on Longman’s hands by June of that year. In any case, he wrote to Cairnes in April of 1872 to say that, as well as rereading and (to our regret) culling old letters, he had been “correcting proofs for new editions” of the Logic (the eighth, which appeared in July) and the Examination (the fourth, our copy-text, which appeared in October).⁴⁷

III.

THE REVISIONS

commenting on mill’s replies in the third edition “to the host of critics” who had assailed the Examination, Bain says, with justice: “The additional scope given to the author’s polemical ability greatly enhanced the interest of the book.”⁴⁸ Indeed the temper, the tone, and to a significant extent the focus of the work were altered by the revisions in the third and fourth editions. As is so often the case, Mill’s own comments in his Preface to the third edition (and those added there in the fourth) give no clear guidance to his rewriting and imply that much less took place than is the actual case. In 1867 he wrote:

And in the fourth edition he calls attention only to the two corrections deriving from Veitch and a reply to Ward (see the discussion below).

The changes were very considerable indeed. Using the crudest of measures, the number of pages,⁴⁹ to give a sense of the amount of change, one finds that the first and second editions are of the same length, 560 pages of text.⁵⁰ The third, however, has 633 pages (an additional 73, or 13 per cent), and the fourth has 650 (a further 17 pages, or 3 per cent). This measure even on its own terms seriously underestimates the amount of addition, for much of the new material—far more than in any other of Mill’s heavily revised works—is in footnotes, set in very small type with minimal leading.⁵¹

Substantive variants. As the account just given would indicate, the second edition was very little revised. Of the total of almost five hundred substantive variants in all editions, fewer than forty occurred in the second edition, almost all of them being very minor revisions of wording in the text. The great bulk of the changes, some 345, or just over 70 per cent, were made in the third edition, and of these nearly one-third were added footnotes or parts of footnotes. The fourth edition accounts for the remaining one hundred odd variants, with an even larger proportion (about two-fifths) being either added footnotes or parts of footnotes (the latter being here more significant than in the third edition, as Mill responded to criticisms of replies he had added in notes in 1867).

For purposes of comparison as well as analysis, one may classify the variants into four groups: (1) major alterations, involving changes of opinion, the introduction of new information, and responses to criticism; (2) changes resulting from the passage of time; (3) qualifications and clarifications of a minor kind, generally involving semantic shifts; and (4) minor changes in syntax, changes entailed by other changes, italicization, terminal punctuation, and merely referential footnotes. In Mill’s other works one finds, as would be expected, a great preponderance of changes of the third and fourth kinds; in the Examination, however, there are as many of the first kind as of the fourth (just over 180 in each case), comparatively fewer of the third kind (120), so typical of Mill elsewhere, and only a handful (8) of the second kind. What may appear strange about this pattern disappears on closer inspection: the vast majority of the type (1) changes (two-thirds of which occur in notes) are responses to critics of a kind rare even in the Logic and the Principles. The paucity of type (2) changes of a simple sort is explained by the relatively short time (seven years) from the first edition to the last in Mill’s lifetime, and by the nature of the text, which is such as virtually to preclude comments that would be affected by the passage of a few years. As to the slightly smaller percentage of type (3) changes, it may be noted that while such changes are found in virtually everything republished by Mill, the greatest volume of them occurs in editions revised in the early 1850s.

The distribution of the changes within the work is informative, but before turning to such questions it is worth citing a few examples to illustrate in general the sort of revision that Mill engaged in, and to call attention to some features that might otherwise not be strongly evident. As indicated above, most of what have been counted as type (1) changes are in response to criticism. As an illustration of those occurring in footnotes, one may cite 32n-3n, added in the third edition like most of the other notes not found in the first edition, where Mill quotes Alexander Fraser in support of his position, as against Mansel and the anonymous reviewer in the North American Review, and goes on to differ from Fraser’s interpretation of Hamilton. The note includes (32^t-t) a variant arising from a type (3) revision (a qualification) in the fourth edition, and concludes with a lengthy passage added in the fourth edition, arising from the continuation of Hamilton’s “Dissertations on Reid” having been called to Mill’s attention (a matter discussed at lxxiii above). This note is typical of others in its length, in its dealing with more than one critic (and issue), and in its containing elements from both the third and fourth editions.

Of the major changes that occur in the text rather than the notes, several may be cited to illustrate different motivations and results. Very few passages were deleted; one of the longest instances (as usual, deriving from the third edition) occurs at 19^e—but in fact the deletion is only seemingly made, for, in revised form, most of the text is used in the long addition, 20^g-g22, which shows Mill responding to the criticisms of John Cunningham and Mansel. (An actual deletion of considerable length will be seen at 189^f.) Another long addition in the text, exceptional in its length, but again typical in dealing with more than one critic (four in this case) and deriving from the third edition, will be seen at 24^m-m32. An interesting example of a more temperate or cautious judgment is seen at 82ⁿ, where (discussing Hamilton’s views of antinomies) Mill originally commented: “I think he has failed to make out either point”, but in the third edition deleted the sentence. Actually this change is related to others occurring later in the chapter (see 88n and 87^c-c), where the justification will be found. (The footnote on 81 is one of the few where Mill mentions what he had said “in the first edition”; actually in these cases—which are not full retractions—the matter appeared in both the first and second editions.)

Most of these examples, as mentioned above, relate to criticism, though seldom does Mill admit to actual mistakes in fact or judgment (in the last example, he refers merely, 88n, to “an over-statement”); there are a few places, however, where he makes—not in a full spirit of repentance—revisions. His controversy with John Veitch, Hamilton’s biographer, to which further reference will be made below, led to Mill’s admission in the Preface to the fourth edition (cvii-cviii below) that he had made two “errors” which, though they did not “affect anything of importance in the criticism” of Hamilton’s interpretation of Aristotle, needed correction. These may be seen in Mill’s last chapter, at 503^k-k and 503n, in the first of which Mill silently added “by the editors,” in response to Veitch’s complaint that Hamilton was not guilty of mistaking the meaning of an Aristotelian term, his editors having searched out a passage to bear out his text. The second of these changes is almost parallel, except that here Mill mentions Veitch and the accusation, and goes on to say that the editors “would have done more wisely by making no reference, than one which so totally fails to support the inference drawn from it” (503n).⁵² Another interesting correction will be seen at 143n and ^p-p: Mill, as always, had assumed that in mentioning Herbert Spencer he would give no offence; Spencer, as always, took offence; and Mill, as always, hastened to apologize—without giving very much away.

Another kind of variant included in type (1) reflects Mill’s work on his other writings. Most evident here are additions bearing a relation to his “Bailey on Berkeley’s Theory of Vision” and “Berkeley’s Life and Writings.”⁵³ The former appeared in Dissertations and Discussions, Vol. II, which Mill revised for its second edition (March, 1867), just before completing his revision of the Examination for its third edition (May, 1867). It is therefore reasonable to assume that the adding in the third edition of 178^a-a was the result of his having just looked carefully over this article, especially if one compares 178^a-a with pages 249-50 and 255 of the article.⁵⁴ Similarly, the later article, “Berkeley’s Life and Writings” (published in November, 1871), was fresh in Mill’s mind when he revised the Examination for its fourth edition (1872), and surely one may assume that his reading for that article is reflected in the additions in 1872 seen at 163^g-g and 230^s-s (with the second, cf. “Berkeley’s Life and Writings,” 456-7). A variant connected in subject matter with these is of a slightly different sort: in the third edition a reference was added 236^z-z to Thomas Nunneley’s On the Organs of Vision, published in 1858, which (given that date) Mill could have mentioned (it would have been apposite) in revising “Bailey on Berkeley” (see, e.g., 263-5 of that article) in 1867, but did not. Nunneley is mentioned, however, in the passage (454-7) of “Berkeley’s Life and Writings” that treats of the same issue.

A regrettably frequent source of confusion for readers wishing to identify or check Mill’s references may be illustrated by 85n, added in the third edition, and modified in three places in the fourth. The note begins: “Mr. Mansel replies (p. 134) . . .”⁵⁵ —but Mill does not here say to which of Mansel’s works he refers (he quotes from four in the Examination). Admittedly, if readers knew that the note had been added in 1867 they would have been likely to guess that Mill was citing The Philosophy of the Conditioned, which appeared in 1866 as a reply to the Examination. Even if that guess had been correctly made, however, the reader would hardly suspect that the final two sentences of the note were added in 1872 (see 85^v-v), and the reference therein (“Mr. Mansel says we do . . .”) is to Mansel’s reply to Mill’s third edition in his “Supplementary Remarks,” in the Contemporary Review for September, 1867. (Compare the long footnote, note, 76n-7n, where Mill in the fourth edition, without notice, sandwiches between two passages of the third edition three sentences, including a quotation, referring simply to Mansel’s “rejoinder”—i.e., again the essay in the Contemporary Review.) Similar problems must have beset careful readers when Mill does not indicate that he is referring to two of McCosh’s works at, for example, 75n, where there is a footnote compounded of 1867 and 1872 passages, in which, again, neither title is given.

Actually Mill was aware (how could he not have been?) of the desirability of indicating that certain passages had been added subsequent to the first publication. He wrote to Augustus De Morgan in 1865: “I have sometimes thought I ought to have some mark for alterations and additions. But one could scarcely give distinctive marks to all the successive strata of new matter, and a mere note of distinction from the edition immediately previous would not answer the [purposes of] those readers who only possess a still earlier one.”⁵⁶ In the third edition of the Examination he in fact made a much less than half-hearted attempt, and, alas, a misleading one, to indicate added footnote material. Of the more than one hundred footnotes or parts of footnotes that were added to the third edition, some ten are parts of notes, and of these four appeared in 1867 between square brackets.⁵⁷ And in 1872, of some twenty parts added to notes, five were placed between square brackets.⁵⁸ Unfortunately, not only is the device used sporadically for parts of notes and not at all for full notes and the text, but also it is unexplained and confusing (there is no distinction, for example, in the fourth edition between additions made in 1867 and those made in 1872). Even assuming that Mill intended to indicate only additions within footnotes that he considered important (and here one would want to challenge his judgment), examination reveals problems. For example, in the footnote that appears below on 71-4, the paragraph on 72n, running “Hardly . . . spare.” and that on 74n, running “The ‘Geometry of Visibles’ . . . truths.” were added in 1867 as the contiguous concluding paragraphs to a note found in the first and second editions; Mill placed square brackets around them. However, in 1872 he added six paragraphs between those two, placed square brackets around those added in 1872, and deleted those around the two added in 1867. So the indication of the earlier addition (an indication he evidently thought important in 1867) is lost, even to anyone trying to understand the device, and such a person would also, at least prima facie, assume that the unbracketed portions of the note all date from the same edition. Another confusing instance is that at 93n-4n, where the square brackets were added in 1872 to a passage introduced in 1867 without brackets; uniquely, the brackets here evidently signal the rewriting that occurred for the fourth edition.⁵⁹

Some of the type (1) changes, it will be noted, entail other changes, which have been counted as type (4). An example is to be found at 154^c-c, where in 1872 an addition to the footnote of a needed explanation of Mill’s judgment that a particular element in Kant’s reasoning is “strangely sophistical” resulted in the deletion of that characterization from the text (154^b).⁶⁰ Similarly, the footnote added in 1867 to 266 (a response to Mahaffy’s argument), entailed (as Mill therein explains) the addition of “persistent” to the text (266^i-i).⁶¹ It need hardly be mentioned that many additions and revisions, especially the longer ones, brought with them referential footnotes, which similarly have been counted as type (4) variants (see, e.g., 24n, which results from 24^m-m32).

Before leaving the type (1) variants, one may mention a few of a minor kind. One of these shows Mill as sharing the frailties of most people: on 386, in an illustrative logical example, he (carelessly?) said, “A dolphin is a fish”; sometime before the third edition the error was caught, and “herring” swam into the dolphin’s place. And finally, the added reference in 1867 to Whately’s Logic at 410^a-a may indicate that he had again looked at that work so important in his mental history—or, of course, someone such as Alexander Bain (then working on his own Logic) may have mentioned the appositeness of the citation.

Type (2) changes, that is, those resulting from the passage of time, being rare in the Examination, may be dealt with briefly. One obvious type, reflecting a changed status (or a change in Mill’s knowledge of status) may be seen at 92^d-d, where “Mr. Calderwood” becomes “Dr. Calderwood” in 1867, and at 164^k-k and 165^l-l, where Ward’s doctorate is similarly recognized in 1872.⁶² The reasons for other changes of this type are less easy to establish: at 116^f a reference to his father changes in 1867 from “Mr. Mill” to simply “Mill”;⁶³ and in 1872 at 216^f, “Professor Bain, of Aberdeen,” loses his institutional identification.⁶⁴ It is also puzzling to see that Mill added an apposite reference (217^h-h) to his Auguste Comte and Positivism (published in 1865) in 1872, rather than in 1867.⁶⁵ The added reference in 1872 to Cazelles’s writings at 250^a-a is easier to explain—at least until one tries (in our case almost in vain) to identify exactly which published (rather than proposed) works Mill intends. An example of the interesting kind of type (2) change found frequently in the Logic is seen at 422^d-d, where Mill mentions, in 1867, that there had been “developments of the doctrine of the Unity of Force” since Hamilton’s death.

Moving to type (3) changes, those involving qualifications, we may begin with one very typical of Mill’s revisions: at 280^g-g, in the second edition, he added the words “appear to” in the passage asserting that Pasteur’s “important experiments . . . appear to have finally exploded the ancient hypothesis of Equivocal Generation. . . .”⁶⁶ Other examples of his continuous search for the precisely correct way of expressing uncertainty are quite common: see, for example, 8^{e-e, f-f, g-g}, where in 1867 “we should know” becomes “we might know”, and “since the new impressions would doubtless be linked with the old” becomes “if the new impressions were linked with the old”. For further illustration of Mill’s habit of mind, see 183^k-k (“unintentional” added before “sanction” in 1867), 188^c-c (“an even” substituted in 1872 for “a much” before “more unqualified manner”, in what may be called a combined precept and example as Mill tries to avoid the imputed sin), 213^b-b (“(as I believe, with nearly all philosophers)” modified to “as I believe (with the great majority of philosophers)”), and 237^a (the removal of “fully” in 1867, in a change whose significance becomes apparent when one reads 237n-8n, also added in 1867). The example at 397^f-f is interesting in that it shows a reversion to an earlier reading of a single word which might be taken (on that ground as well as on the ground of sense) as a typographical error (and is so questioned in the variant note), but which could represent a hesitancy over a legitimate choice of words (“subject” was altered to “object” in the second edition, with “subject” being restored in the third).

A type (3) change of a significant kind, representing a search after more precise expression of a concept, may be seen at 220^i-i, where Mill says (in 1865) of Brown and others who held the psychological theory: “Their argument is not, as Sir W. Hamilton fancied, a fallacious confusion between two meanings of the word length, but an identification of them as one”, and (in 1867), substituting a semi-colon for the last comma, replaces the last clause with: “they maintain the one to be the product of the other.”⁶⁷ Compare, as a type, 141^m-m, where the original wording, “the time at which memory commences”, was altered (in the fourth edition) to the more accurate, “the time to which memory goes back”. Another slight example of the kind found in larger number in works more often revised by Mill is seen in the double change at 178^c-c: here Mill originally wrote, “there is in our perceptions”; in the second edition he altered the wording to “there is involved in our perceptions”; and he settled, in the third, for “there is concerned in our perceptions”.

One final illustration of type (3) changes points again to Mill’s frequent tempering of judgments in what, at this stage in the controversy and in his career, cannot be seen as mere caution. At 480^d-d he first published his opinion of a blunder in this form: “If Sir W. Hamilton could think so, his ignorance of the subject must have been greater than can be imputed to any educated mind, not to speak of a philosopher.” In 1867, rewriting of the first part of the sentence produced a still stern, but less particular and insulting condemnation: “to think so would require an ignorance of the subject greater than can be imputed to any educated mind, not to speak of a philosopher.”

There is no need to dwell on the type (4) variants, which on the whole reflect the sort of revision in which anyone engages who tries to make syntax more transparent and emphasis more obvious. A few, however, may be cited, just to suggest the effect of such fine tuning, and to show that some are not entirely trivial. At 175^g-g the change (one of the rare second-edition variants) from “nothing different in it from his own” to “nothing in it different from his own” clearly makes the sentence easier to read. Throughout the volume Mill habitually uses the first-person singular, and seems to have been more careful than many of his “cotemporaries” in saving the first-person plural for editorial (as well as normal) usage: it is therefore slightly surprising to find “we” at 136^i-i, but not at all surprising to see that “I” replaced it in 1867. Simple removal of unnecessary emphasis is seen at 6^a-a and 7^d-d, where italics were removed from “outside” and “something” in 1867. One variant of moot significance—it could even be considered a type (1)—is found at 458^r-r, where “On the theory of Necessity (we are told) man cannot help acting as he does” is modified by the perhaps trivial and perhaps important insertion of “a” before “man”.⁶⁸

A major problem for editors (though not for most more fortunate folk) lies in deciding which changes in a text should be seen as minor variants and which as printer’s errors.⁶⁹ Some examples where the latter choice was made are 119.31 and 35,⁷⁰ 348.11,⁷¹ 382.14,⁷² 478.21⁷³ and 483.19.⁷⁴ Examples of the former choice, that is, where the evidence and/or sense suggest that a variant reading is useful (even when, in some cases, a typographical error is almost certain), are 42^q-q, 306^a-a,⁷⁵ 289^b-b, and 382^f-f.⁷⁶ It is virtually certain that the printer misread Mill’s hand in places, but in general it seems wisest to adopt the conservative principle of retaining what appears in the printed text, except where there is strong evidence; some such cases of caution are revealed if one looks at the collation of Mill’s quotations with their sources, as at 203n.5 and 8, where the compositor read (and we print) “any,” where O’Hanlon, whom Mill is quoting, has “my.” A pair of typographical errors deserve mention because they signal the existence of two states of Gathering K in the fourth edition: in the correct state, at 103.35 the reading is “But if what I am told”, and at 104.19, “Is it unfair . . .?”; in the incorrect state (probably the second, resulting from the forme being pied), the readings are “But if what am I told”, and “It is unfair . . .?” (There are other, non-substantive, indications of resetting in the gathering.)

Accidental variants. The pattern of changes in punctuation does not match that of the substantives, for the largest number (over 190) occurs in the second edition, the great bulk being, as expected, the deletion of a comma or a pair of commas (72 cases, the most numerous kind of change overall) or the addition of a comma or a pair of commas (34 cases). Considering only types of change where there are at least ten instances, one finds in the second edition twenty-six places where a semi-colon was substituted for a colon, and nineteen places where the reverse change occurs. In fourteen instances Mill reduced an initial capital letter to lower case.

In the third edition there are about 170 changes in punctuation and initial capitalization, markedly fewer in relation to the accidental and substantive changes in the second and fourth editions, when judged by the pattern in Mill’s other works. The reason may be that none of his other heavily revised writings received their most thorough reworking this late in his life, or without his wife’s assistance, and that, with so much of the substantive revision consisting of added footnotes, Mill scrutinized the text less carefully—or it may be that, by his judgment at least, the second edition was quite well punctuated—or, indeed, the explanation may lie, at least in the main, in the habits and predelictions of the compositors of this and other of Mill’s works. In any case, the third edition reveals again as the most frequent changes the addition (52 instances) or deletion (51 instances) of individual or paired commas; next most frequent are the lowering of initial capitals (19 instances), though here initial letters are raised eleven times; and in ten instances a semi-colon replaces a colon. In seven instances of various kinds the changes have been judged to be typographical errors.

The fourth edition reveals some 150 changes of these kinds, the most interesting fact about them being that forty are of the sort we have considered as typographical errors (and so in these cases we have in our text adopted the reading of the third edition). Of the total, the addition (27 instances, four read as typographical errors) and deletion (62 instances, 19 read as typographical errors) of individual or paired commas again predominate; continuing the general pattern of lightening punctuation, in fifteen cases (three seen as typographical errors) semi-colons replace colons; in raw scores, raising and lowering of initial letters tie with ten instances each, but nine of the latter, as against two of the former, appear to be typographical errors.⁷⁷

The spelling changes provide, as is usual in Mill’s texts, more opportunity for speculation than grounds for judgment, especially in the absence of manuscripts and proof. The most common alterations are from “s” to “z” (and the reverse) in verbals, and of initial “i” to “e” (and the reverse). Of the first of these, the treatment of “cognize” (and its cognates) will illustrate Mill’s (or someone else’s) indecision: in the third edition, in one instance “s” becomes “z”, while in another “z” becomes “s”; in the fourth edition, in three cases “s” becomes “z”, while in two the reverse change occurs; in two passages added to the text, one in the third and one in the fourth edition, the former uses the “z” form (which is retained in the fourth edition), while the latter uses the “s” form. Changes from “e” to “i” (and the reverse) include six cases (four in the second edition, two in the third) where “enquiry” (or one of its cognates) becomes “inquiry”, two (one in the second edition, one in the fourth) where cognates of “inclose” become “enclose”, and the alteration in the third edition (one instance each) of “intangle” and “indorse” to “entangle” and “endorse”. As to the vexing question of final single or double “l”, in four cases (on the same page) in the second edition “recall” became “recal”, and the same change is found twice in the third edition and once in the fourth—but also in the fourth the four “recal”s of the second edition reverted to the “recall” of the first. (The one use of “foretel” persists through all editions, and “dispel”, added in the third edition in one place, remains in the fourth.) There seems no clear guidance as to whether or not Mill preferred a hyphen after the prefix “co”, except in “coexist” and its cognates, where the clearly dominant form is without the hyphen; also it seems doubtful to assume that he came to prefer “phenomenon” to “phænomenon”, because, although the former is adopted once in each of the third and fourth editions where a change occurs, as well as in more than ten passages added in those editions, the latter form persists. All in all, it seems wise to conclude that many, though not all, of the changes reflect the preferences of compositors rather than of Mill.

Mill’s references and sources. As the Bibliographic Index (Appendix D) reveals, there are direct or indirect references to about 190 works in the Examination, and over 80 references to persons not specifically as authors. Of the cited works, nearly 60 per cent are quoted, the bulk of the quotations coming, of course, from the writings of Hamilton. There are references to, and quotations from many of, twenty-two books or reviews prompted at least in part by Mill’s attack on Hamilton, and Mill refers to five of his own writings (usually because they were mentioned by his critics). One interesting finding is that of the works Mill cites when controverting Hamilton’s view of contrariety, having, he says, “only looked up the authorities nearest to hand” (412-13), the London Library has copies of eight which he had known from youth.⁷⁸

For the most part his treatment of his sources is fair, and transcriptions reasonably accurate, but of course his judgments were polemical, and much resented by members of the other “school” of philosophy. A good deal of the argument is carried on, as is common in the genre, by quotation and counter-quotation, so the proportion of quoted matter is much higher here than in Mill’s other major works.⁷⁹ Mill made a genuine attempt to answer his critics, but he was as little sympathetic to some of them as they were to him, and so it is misleading to estimate either the strengths or the weaknesses of his opponents (or even his allies) by his citations in the Examination. A few specific instances may be mentioned, partly at least for their curiosity.

The changes Mill made in the fourth edition as a result of Veitch’s criticisms have already been touched upon, but the matter merits a few further words, for Mill chose to ignore most of what Veitch had to say. The justification is in a letter to Bain:

Veitch (who was using the first and third editions of the Examination, with page references to the third) attacked Mill for alleging that Hamilton’s philosophic positions were conditioned by his unreasoned acceptance of the doctrine of free will, and that he bribed his pupils to accept metaphysical dogmas “by the promise or threat” that they afford the only valid support for “foregone” conclusions.⁸¹ These remarks Mill ignored, as he also passed by Veitch’s admission that Mill was in the main right in suggesting that Hamilton lacked (in Veitch’s phrasing) “the historical imagination as exercised in philosophy,” though Mill notices Veitch’s claim that Mill was completely wrong in imputing to Hamilton a weakness in perceiving (Mill’s words) “the mutual relations of philosophical doctrines.”⁸² This latter question is gone into more thoroughly by Veitch in his appended Note C, where he examines and attempts to refute Mill’s expositions of Hamilton on Hume, Leibniz, and Aristotle (to only part of the last of which did Mill respond, in the changes alluded to above).⁸³ The “vinegar” of Veitch’s attack is most evident in his Note A, concerning Hamilton on Cousin’s view of the Infinite and Absolute, where, in language stronger than Mill’s about Hamilton, Veitch refers to Mill’s “gross, even ludicrous, misrepresentation of Hamilton’s doctrines,” and says, in a classical example of the rhetorical device of occupatio, that there is no need for further rebuttal than that found in Mansel’s “admirably clear, acute, and powerful exposure of Mr Mill’s misconceptions” in his Philosophy of the Conditioned.⁸⁴ Given Veitch’s special acquaintance with Hamilton’s Lectures (he was one of Hamilton’s students),⁸⁵ it is interesting to find him assailing Mill for treating them as of equal value with Hamilton’s “deliberate writing”;⁸⁶ it is even more interesting to find Veitch nonetheless using the Lectures in an attempt to refute Mill on a substantive issue,⁸⁷ and then showing no hesitation in bestowing fulsome praise on them in other contexts, and going so far as to devote his Note B to citing the high opinions of the Lectures expressed in the United States.⁸⁸

The emotional disadvantages of engaging in this kind of controversy are illustrated by the rather unusual reaction of Patrick Proctor Alexander to Mill’s ignoring his riposte to Mill’s response to his attack. Alexander, in what Mill accurately characterized as a “rollicking style” (460n), assaulted the Examination in Mill and Carlyle. An examination of Mr. John Stuart Mill’s doctrine of causation in relation to moral freedom. With an occasional discourse on Sauerteig, by Smelfungus (Edinburgh: Nimmo, 1866)—the title itself giving clear enough indication that Alexander inclined to the Carlylian side of the conjunction. In the third edition Mill responded at some length (see the citations in the Bibliographic Index), but not very much to the satisfaction of Alexander, who replied with Moral Causation; or, Notes on Mr. Mill’s notes to the chapter on “freedom” in the third edition of his “Examination of Sir W. Hamilton’s Philosophy” (Edinburgh: Nimmo, 1868), the rollicking introduction to which reveals (7) that others had called his style “disgustingly flippant.” “The success of this work,” Alexander later commented, “was, sooth to say, not much; I am not aware that any one ever bought or read it; and the notices of it in the press were few, slight, and for the most part, I rather think, contemptuous.” He had, however, sent a copy to Mill, and anticipated a reply in Mill’s fourth edition—but no such reply was there! Alexander therefore prepared a second edition of his Moral Causation (“revised and extended”), which he planned to issue so that Mill could reply in a fifth edition. But again frustrated, in this case by Mill’s death, Alexander issued his second edition (Edinburgh and London: Blackwood, 1875), in the Preface to which (as well as the remark about the success of the first edition quoted above) he includes a complaint against Mill’s having said in his fourth edition that only Mansel and McCosh had published rejoinders to the replies in the third.⁸⁹ In this context it should be noted that Mill objects (cvii) to McCosh’s assumption that criticisms unrefuted are triumphant; he calls attention to the fact that the subject of the work is the philosophy of Hamilton, not McCosh contra Mill.⁹⁰

Since anyone who has attempted to follow the intricacies of these controversies with sympathy and understanding may well have felt a heaviness of spirit, I may be forgiven the mention of one other curiosity. The perceptiveness of the rather taciturn John Grote (younger brother of George) might well lead one to anticipate valuable comments on the Examination in his Exploratio Philosophica.⁹¹ But all one finds is the following example of scholarly eccentricity, which does not permit of condensation:

Effects of the revisions. The fourth edition—or, more truthfully, the third, where most of the changes appeared—of the Examination is, in tone as well as length, a markedly different work from the first edition, even though virtually nothing was removed. It is, I believe, both unnecessary and unwise to comment extensively on the different ways in which the two versions affect a reader, but a few comments may assist an understanding of the controversial circumstances. Alan Ryan comments in his Introduction on the two kinds of material in the Examination: on the one hand, and most extensively, an exposition and criticism of Hamilton’s (and to a lesser extent, Mansel’s) views; on the other, an exposition and defence of Mill’s (usually) countering views. While the two cannot be exactly isolated, for Mill almost necessarily interweaves his views with his criticism of Hamilton, Chapters xi, xii, and xiii are devoted to Mill’s account of his own psychological theory, which is compared not in these chapters, but in the surrounding ones, to the views of Hamilton and his school. Also Chapter xxvi, on the Freedom of the Will, contains a good deal of exposition and defence of Mill’s view. One may consider as well the chapters where Mansel receives most attention, especially vii, “The Philosophy of the Conditioned, as Applied by Mr. Mansel to the Limits of Religious Thought,” and xiv, which deals with Mansel’s as well as Hamilton’s treatment of Associationism, as being specially related to Mill’s views.⁹²

So, in the crudest terms, of the twenty-eight chapters (which vary widely in length), six may be considered as most relevant to a consideration of Mill’s direct presentation of his ideas, and all but one of these appear in the first half (measured in chapters) of the work. In fact, measured in pages, in the first two editions these fourteen chapters occupy 270 of the total 560 pages, or 46 per cent. In the third edition these fourteen chapters come to 326 of 633 pages, or 51.5 per cent, and in the fourth to 340 of 650, or 52.3 per cent, not in themselves very startling increases, except that they account for 70 additional pages, leaving only 20 for the chapters of the latter half. When one looks at the five chapters in the first half specially relevant to Mill’s views, the point is more clearly made: 42 of the 70 pages added to the first fourteen chapters (33 pages in 1867, 9 in 1872) appear there.⁹³ (And of the 20 added in the latter half, 11 appear in Chapter xxvi.) Furthermore, two other chapters in the first half, devoted in the first edition to critical examinations of Hamilton’s views on the relativity of human knowledge (Chapter iii) and the Philosophy of the Conditioned (Chapter vi), were greatly expanded for the third edition, and further enlarged for the fourth, to accommodate some of Mill’s strongest statements of his views countering those of Hamilton’s defenders. In these two chapters (both, it will be noted, again in the first half of the work) a further 26 pages appeared by the fourth edition, and so, adding these to those already accounted for, 68 pages of the 70 added in the first half between the first and fourth editions appear in chapters having to do particularly with Mill’s own views and his arguments against the major metaphysical positions of Hamilton and the application of those views by Mansel. (And Chapter xxvi, which is related to these same matters, accounts for slightly more than half of the additions to the second half.) One should also recall the point made above, that a majority of the additions come in footnotes that consist of Mill’s defence or counter-attack against critics who most frequently are assailing Mill’s views rather than supporting Hamilton’s.

Without going into even more painful games of numbers, one may, I believe, accept certain conclusions about the overall rhetorical and tonal effects of the revisions, and make at least suggestions concerning what seemed important to Mill about his vocal readers’ reactions. In the first version, less than half the work was given over to the exposition of Hamilton’s and Mill’s countering views on metaphysics and psychology; the larger part dealt mainly with other aspects of Hamilton’s thought (see Mill’s explanation of the divisions on 301; cf. 109, 417, 430, 470), most particularly with his logical speculations. By and large, almost no one took up the challenge—the challenge is certainly there, for the polemic is very strong in the latter half—to defend Hamilton on logic, or mathematics, or any other special topic, and so the latter half of the work finally remained (with the exception of Chapter xxvi) much as it had been in the first edition. (The interesting comments in the concluding chapter on Hamilton’s personal qualities do little to affect the tone.) But many a critic seized metaphysical and theological cudgels to belabour Mill—not even here, in general, to defend the corpus of Hamilton—and Mill, not without some selection of ground where response would be, in his view, most telling, took the field of their choice, that is, the areas covered in the first half of the work. There is nothing odd in these reactions, of course, but they do suggest that it was not Sir William Hamilton who attracted the critics’ attention, but the battle between the two philosophies, and the way it was being waged by the active combatant, Mill. Surely it may, at the least, be surmised that here lies the explanation for what has appeared odd to many, Mill’s choice of Hamilton as a subject for what is, after all, his third longest work, and one on which he bestowed much labour. He was looking for a fight, and Hamilton (as he discovered during his careful study) provided both issue and occasion. Organizing the work as he did—and exception can be taken to the details of his dispositio⁹⁴ —he called attention to what he considered most important, and (perhaps) most provoking. The event proved him right, for not only did the argument centre on the issues between intuitionism and empiricism, but it was a clamorous one, more immediately intense than that aroused by any of his other works. The Examination, whatever the modern view, passed its own test.

IV.

THE PRESENT TEXT

as throughout the Collected Works, the copy-text is that of the final edition in Mill’s lifetime, in this case, the fourth, 1872. It has been collated with the three previous editions, and the few manuscript fragments reproduced in Appendix A. Substantive textual changes among the editions are recorded, substantive here meaning all changes except spelling, initial capitalization, word division, punctuation, demonstrable typographical errors, alterations in the form and style of references, and such printing-house concerns as type size.

Our goals are (a) an accurate text as little interrupted by editorial apparatus as is consistent with (b) the immediate reconstruction of earlier versions without separate instructions for each variant, and (c) the minimum number of levels of type on the page. This minimum number is three: the text of the fourth edition; in slightly smaller type, Mill’s own footnotes and referential footnotes added (in square brackets) by the editor; and in still smaller type, footnotes giving the variant readings. In the text itself, the usual indicators (*, ^†, etc.) call attention to Mill’s footnotes; editorial notes of reference are signalled by the same indicators (in separate sequence) enclosed in square brackets. Small italic superscript letters, in alphabetical sequence (beginning anew in each chapter) call attention to variant readings (and, in the seven cases where manuscript fragments occur, to their existence). These variants are of three kinds: addition of a word or words, substitution of a word or words, deletion of a word or words. The illustrative examples below are chosen for their ease of presentation and reference, not for their significance.

Addition of a word or words: see 83^p-p. In the text, the word “invariably” appears as “^pinvariably^p”; the variant note reads “^p-p+67,72”. Here the plus sign indicates that the word was added; the following numbers indicate the editions in which the added word appears. The editions are indicated by the last two numbers of their publication dates, with superscript numerals to distinguish between the first and second editions, both of which appeared in 1865: that is, 65¹=first edition (1865), 65²=second edition (1865), 67=third edition (1867), and 72=fourth edition (1872), the copy-text. The only exception is that one change of “a” to “an” (before “useful” at 471.11) and three changes of “an” to “a” (before “hyperphysical” at 190.9, and twice before “hypothetical” at 410.12 and 16) are not recorded. If a variant occurs within a quotation, and the earlier version (i.e., that in the variant note) is the reading of the source from which Mill is quoting, the word “Source” precedes the edition indicators in the variant note (see 382^f-f). (If the reading in the text, as opposed to that in the variant note, is the same as that of the source, no such indication is necessary.) If the quoted text varies from the source, but does not vary among editions, there is no variant note (the variant reading is given, however, in Appendix D, the Bibliographic Appendix: see, e.g., the entry for 242n.3 under Abbott’s Sight and Touch, 521, where the “then” in Abbott does not appear in any of Mill’s editions).

Placing the example (83^p-p) in context, then, the interpretation is that in the first and second editions the reading is “as it is interpreted”; in the third edition (1867) this was altered to “as it is invariably interpreted”, and the reading of the third edition was retained (as is clear in the text) in the fourth edition (1872), the copy-text.

When the addition is a long one, the second enclosing superscript may appear several pages after the first one; to make reference easier, the superscript notation in the footnote (which appears on the same page as the first superscript) will give the page number where the variant concludes (see, e.g., 63^g-g65).

Substitution of a word or words: see 63^f-f. In the text the words “one of the chief sources” appears as “^fone of the chief sources^f”; the variant note reads “^f-f65¹,65² the chief source”. Here the reading following the edition indicators is that for which “one of the chief sources” was substituted; again putting the variant in context, the interpretation is that in the first and second editions (65¹ and 65²) the reading was “and be the chief source of the reputation”; in the third edition this was altered to “and be one of the chief sources of the reputation”, and this reading was retained (again as is clear in the text) in the fourth edition.

Deletion of a word or words: see 75^g. In the text, a single superscript ^g appears centred between “but” and “could”; the variant note reads “^g65¹,65² it”. Here the word following the edition indicators was deleted; again putting the variant in context, the interpretation is that the reading in the first and second editions was “but it could not realize”; the word “it” was deleted in the third edition, and the reading of the third was continued in the fourth edition.

Passages changed more than once. Here two methods are used. In the cases, rare in the Examination though common in Mill’s other lengthy works (which went through many editions over an extended period of time), when a few words were altered and then altered again, the method followed is that illustrated at 38^k-k. Here the text reads “^juncaused, and is therefore most naturally identified with the^j”; the variant note reads “^j-j65¹, 65² a] 67 the”. The interpretation is, that in the first and second editions the reading, in context, was “In this signification it is synonymous with a First Cause.” In the third edition the sentence ending was altered to “with the First Cause”; and in the fourth edition (as is evident in the text) to “with uncaused, and is therefore most naturally identified with the First Cause.”

The other method is used for changes within lengthy passages added subsequent to the first edition. Most of these, in the Examination, occur within added footnotes (discussed below), but examples in the text will be found at 26^n-n to ^q-q, where within a long addition (^m-m, which runs from 24 to 32) there are, all within one sentence, later changes indicated. Passage ^m-m was added in the third edition, but the wording, in that sentence, was altered in the fourth edition. For example, in 1867 it began: “Indeed, the very fact that Sir W. Hamilton thinks it possible for philosophers to discriminate . . .”; in 1872 the wording became: “Indeed, the discrimination which Sir W. Hamilton thinks it possible for philosophy to make. . . .” In these cases, within the superscript letters indicating an added passage, there will be other superscript sets indicating other changes; the variants are listed separately, and in the order of their appearance in the text.

Variants in Mill’s footnotes. By far the most common type is the note added in full (the great majority of them in 1867), as, for example, at 21n, where the first note begins: “[67] This is essentially . . .”; the editorially inserted “[67]” indicates that the note was added in the third edition, and was retained (as is evident) in the fourth. Many of these are simply referential, deriving from variants in the text proper, but their addition is always separately signalled in this way.

Changes within notes are treated in the same manner as changes in the text: see, e.g., 29^g-g, where a passage was added in 1872; 29^r-r, where a substitution was made, again in 1872; and 34^a, where a clause was deleted, once more in 1872.

Prefaces. No preface appeared in the first or second edition. That to the third edition was reprinted (still entitled “Preface to the Third Edition”) in the fourth, with a substantial addition. Mill not having indicated that this matter was added in the fourth edition, we have treated it as a variant in the usual way.

Other textual liberties. The following changes are all silently made in the text, except as specifically indicated. Textual emendations, including typographical errors, are listed in Appendix B below, with a note explaining their choice and treatment.⁹⁵ (Typographical errors found only in one or more of the first three editions are not listed.) Long quotations have been set in smaller type; the quotation marks found in such quotations at each line in the left margin of the original editions have been removed. (It may be noted, as will be seen in Appendix B, that in several instances these quotation marks led to typographical errors when the text was reset.) Within these quotations, Mill sometimes used quotation marks and round brackets to signal interpolations; we have deleted the quotation marks and substituted square brackets. As mentioned above, the square brackets that Mill occasionally used to indicate matter added in footnotes have been replaced by variant indicators. Mill’s placing of footnote indicators was rather eccentric in this work; we have, wherever it was possible without causing confusion, moved them to the ends of quoted passages. Infrequently the result is that two of his notes have been combined. Also, more infrequently, where Mill gave a single reference for a very lengthy quotation from which he had omitted a considerable passage, we have split his one reference into two. Indications of ellipsis in quotations have been standardized to three dots plus, when required, terminal punctuation. A few trivial alterations in printing style have been made, such as the removal of dashes when combined with other punctuation in introducing quotations and references. The running heads have been altered to suit this edition. When necessary, Mill’s references to sources have been amplified and corrected (the corrections are listed in Appendix C below), with all added information being placed in square brackets, as are all editorial references. These last are also signalled by indicators in square brackets, as mentioned above.

V.

APPENDICES

Appendix A gives the readings (with explanatory and variant notes) of the few manuscript fragments that survive.

Appendix B lists the textual emendations with the original readings. The headnote gives the general justification; individual items there entered give, when necessary, the special justifications.

Appendix C gives the original and emended readings of Mill’s references that have been silently corrected in his footnotes.

Appendix D, the Bibliographic Appendix, lists all persons and works referred to or quoted in the Examination, except mythical persons and those simply used as place-holders in logical examples. Substantive variants between Mill’s quotations and his sources are entered, both to correct misquotations and to provide contexts for partial quotations. Because this appendix includes all references to persons and books, it is in effect an index to names and titles, which are therefore omitted from the Index proper.

The Index has been prepared by Dr. Bruce L. Kinzer.

ACKNOWLEDGEMENTS

for permission to publish manuscript material, we are indebted to the Columbia University Library, the Houghton Library of Harvard University, to the Yale University Library, and to the National Provincial Bank (literary executors and residual legatees of Mary Taylor, Mill’s step-grand-daughter). Our deep gratitude is once again cheerfully offered to the staffs of the British Library, the Somerville College Library, the University of London Library, the University of Reading Library (and especially its Archivist, Mr. J. A. Edwards), the British Library of Political and Economic Science, the London Library, the University of Toronto Library, the library of the Pontifical Institute of Mediæval Studies, the libraries of Knox, Regis, St. Michael’s, Trinity, and Wycliffe Colleges, Toronto, and never least, the Victoria University Library. To the members of the Editorial Committee, especially Jean Houston and R. F. McRae, to the copy-editor, Rosemary Shipton, and the editorial, production, and printing staff of the University of Toronto Press, my never failing, though not always expressed, thanks for unstinting co-operation. Among others to whom credit is due, and no discredit should accrue, are Father J. L. Dewan, Mr. Charles P. Finlayson of the Edinburgh University Library, Professor Daniel De Montmollin, Professor Joseph Hamburger, Professor Hugh R. MacCallum, Professors E. Jane and Michael Millgate, Professor Emeritus J. R. O’Donnell, the Reverend J. Owens, Mr. H. Russell of the Belfast Public Libraries, Professor C. A. Silber, Professor F. E. Sparshott, Professor Jack Stillinger, and Professor J. R. Vanstone. In a very real sense the editing of the volume is the work of my colleagues on the Mill project, where good spirits, co-operation, and industry have made the time seem short and be pleasant: Marion Filipiuk, Bruce Kinzer, Martin Kreiswirth, Judith LeGoff, and Rea Wilmshurst. Lady Hamilton, among her other duties, sat up through the nights till the northern dawn transcribing the lecture notes which Sir William was writing for the next afternoon; that my wife did not do the like for me (nor I for her) might suggest a number of conclusions, but I believe that her generous aid to this volume would not have been forthcoming had she (or I) done so.

[1 ]The work is identified in Mill’s bibliography as “An Examination of Sir W. Hamilton’s Philosophy, and of the Principal Philosophical Questions discussed in his writings. 8vo Volume, first published in 1865.” (MacMinn, 96.) In his library, Somerville College, Oxford, is a copy of the first edition, without corrections or changes, and for the most part unopened, and a copy of the two-volume American edition, New York: Holt, 1873.

Fifth and sixth London editions appeared in 1878 and 1889 (the latter also Longmans, New York) while American editions appeared in 1866, 1868 (both Spencer of Boston), 1873, and 1884 (both Holt of New York). A French translation was published in 1869 (Paris), though, surprisingly, there seems to have been no German translation until 1908 (it was not included in the twelve-volume Gesammelte Werke edited by Theodor Gomperz).

[2 ]In 1865, in addition to the two editions of the Examination, Mill published the periodical and first book editions of Auguste Comte and Positivism, the third edition of Considerations on Representative Government, the sixth editions of both the Logic and the Principles, and People’s Editions of the Principles, On Liberty, and Representative Government.

[3 ]See David Masson, ed., The Collected Writings of Thomas De Quincey, 14 vols. (Edinburgh: Black, 1889-90), V, 308.

[4 ]See Francis E. Mineka, ed., The Earlier Letters of John Stuart Mill, Collected Works, Vols. XII and XIII (Toronto: University of Toronto Press, 1963), XII, 173 (2/8/33) (henceforth referred to as EL, CW, with volume and page numbers); and, for Carlyle, Alexander Carlyle, ed., Letters of Thomas Carlyle to John Stuart Mill, John Sterling, and Robert Browning (London: Unwin, 1923), 61 (18/7/33), and 78 (28/10/33). Mill may also have known that Captain Thomas Hamilton, author of Man and Manners in America (1833), was Sir William’s brother. (Mill reviewed a review of Captain Hamilton’s book in his “State of Society in America,” in Essays on Politics and Society, Collected Works, Vols. XVIII and XIX [Toronto: University of Toronto Press, 1977], XVIII, 91-115.)

[5 ]See John Veitch, Memoir of Sir William Hamilton, Bart. (Edinburgh: Blackwood, 1869), 128 (letter of 8/7/34), and 175 (letter of 26/11/34).

[6 ]EL, CW, XIII, 528 (to Austin, 7/7/42). In fact, no review appeared.

[7 ]The order and form of publication of Hamilton’s works, and the order in which Mill read them, led to some confusion, as will be seen below.

[8 ]See the account in the Textual Introduction, Essays on Ethics, Religion, and Society, Collected Works, X (Toronto: University of Toronto Press, 1969), cxxii ff.

[9 ]Autobiography, ed. Jack Stillinger (Boston: Houghton Mifflin, 1969), 160-1. The former article was published in Fraser’s in February, 1862; the latter in the Westminster in October, 1862.

[10 ]Ibid., 161.

[11 ]Bain, John Stuart Mill (London: Longmans, 1882), 118 (cf. Francis E. Mineka and Dwight N. Lindley, eds., The Later Letters, Collected Works, Vols. XIV-XVII [Toronto: University of Toronto Press, 1972], XV, 746; subsequently referred to as LL, CW, with volume and page numbers). Bain adds that Mill “chose the Westminster when he wanted free room for his elbow.”

[12 ]See Bain, ibid., letter of December, 1861 (LL, CW, XV, 752): “. . . I have given up the idea of doing it in anything less than a volume.”

[13 ]Presumably Mill is using the short title, and not implying that he had ignored the rest of Discussions on Philosophy and Literature, Education and University Reform.

[14 ]The matter is called by Mansel “Supplementary Part, to complete Former Editions,” and includes a “Postscript” (989-90), dated 23 August, 1862, explaining (inadequately) the additions; these consist of the remainder of D***, and E-U (there is no J), U*, V-Y (all on 915-88), and “Addenda” (989*-91*).

Another confusion in Mill’s mind, excusable only by the actual dates of publication of Hamilton’s writings, may be seen in the variants at 163, when Mill in the first edition had Hamilton quoting, in his paper on Brown (1830), from his Lectures, which he did not begin to deliver until 1836. But the Discussions (including the paper on Brown) were published in 1852, and Mill may not have noticed when Hamilton began his lecturing.

[15 ]Autobiography, 161.

[16 ]Ibid., 161-2.

[17 ]Ibid., 163.

[18 ]Bain, John Stuart Mill, 118 (LL, CW, XV, 752). Mill added to the sixth edition (1865) of his Logic eleven footnoted references to the Examination (cf. the Textual Introduction, A System of Logic, Collected Works, Vols. VII and VIII [Toronto: University of Toronto Press, 1973], VII, lxxxv and n.).

[19 ]LL, CW, XV, 763. The letter continues: “But I do not feel properly equipped for such a piece of work until I have read your account of Plato, in which I expect to find much new and valuable thought on the great problems of metaphysics.” Though he saw Grote’s Plato in manuscript, it was not in fact published until 1865, too late to be of use for the Examination.

Grote, replying to the letter and refusing Mill’s invitation that he join Helen Taylor and himself on a trip to Greece (which its historian never visited), added: “Your intimation of what you had been doing about Sir W. Hamilton’s works was still more interesting [than a passage in Lucian mentioned by Mill], as it holds out to me the hope that you may one of these days revert to those higher speculative and logical subjects with which he busies himself.” (Harriet Grote, The Personal Life of George Grote [London: Murray, 1873], 257.) When the Examination appeared, Grote wrote to say: “it has completely answered my expectations, and that is saying as much of it as I can say. It is full of valuable expansions of the doctrines more briefly adumbrated in your Logic, and of contributions to the most obscure and recondite expositions of Psychological Science. . . .

“I am certainly very glad that poor Sir W. H. did not live to read such a crushing refutation. It is really so terrible, that I shall be almost pleased if either Mansel or T. S. Baynes are able, on any particular points, to weaken the force of it, and make something of a defence.” (Ibid., 275, letter of June, 1865.)

On 20 November, 1865, while writing his review of the Examination, Grote praised in particular Mill’s treatment of Matter as a Permanent Possibility of Sensation, and his vindication of the derivation of belief in coexistent parts of extension from successive conscious phenomena of motion; here also he notes his pleasure at being able to record his homage to James Mill. (Yale University Library; partly printed in ibid., 278.) Grote’s interest in Hamilton continued: not only did he review the Examination in the Westminster, saying some complimentary things about Hamilton, but also sent Bain a paper on “what Sir William Hamilton says in reference to Aristotle’s views of common sense,” used by Bain in his Mental and Moral Science (1868). (Ibid., 290, letter to Bain of October, 1867.)

[20 ]LL, CW, XV, 809 (14/12/62). The “something” is undoubtedly what became part of Chapter xxi below.

[21 ]John Stuart Mill, 119. Cf. n.30 below.

[22 ]LL, CW, XV, 817 (7/1/63) and 836-7 (13/2/63), both to Bain.

[23 ]Ibid., 866.

[24 ]Ibid., 889.

[25 ]Ibid., 900. For the use made of Bain, see 216-19, 226-7, and 231-6 below.

[26 ]Ibid., 907.

[27 ]Autobiography, 163.

[28 ]John Stuart Mill, 119. See also Mill’s response (124n) to the defence by the “Inquirer” of inconsistencies in Hamilton. It is at the least ironical that Mill himself has been so much assailed for inconsistencies; of course, no one escapes hanging, if not on this charge, then on its opposite, purblind single-mindedness. And some critics wish Mill were alive to answer other of his critics.

[29 ]LL, CW, XV, 901-2. Mill was undoubtedly right in his concluding conjecture (and cf. the regret he expresses in his “Introductory Remarks,” 2-3 below); Veitch comments, understating the case, that Hamilton “was fond of controversial writings, and enjoyed the learned railings of the Scioppian style” (46).

[30 ]LL, CW, XV, 901-2. Such insights as we have into Mill’s habits of composition, being rare, are worth citing. See, for example, Mill’s letter to Bain of 7/1/63, where he mentions going “deliberately through the whole writings of Hamilton, writing down in the form of notes, the substance of what I as yet find to say on each point. This will make it comparatively easy to write the book when I have finished the preparatory work.” (LL, CW, XV, 816.) See also a footnote added in 1867, where Mill (presumably ironically) thanks Mansel for reminding him of two passages he would “not have failed to quote” in the first edition, if he “had kept references to them” (22n). What he sometimes did (as did his father) was to list page numbers in the backs of books, presumably to return to them later to make notes; there are surviving only a very few (and none of them here relevant) of what must have been voluminous copied quotations. There is no evidence that (here unlike his father) he kept a Commonplace Book containing quotable passages.

[31 ]He returned to London in mid-February, went back to Avignon in April, travelled back to London in June, stayed there until early September, and then passed the rest of 1864 and most of January, 1865, in Avignon.

[32 ]LL, CW, XV, 926, 929.

[33 ]Ibid., 945 (26/6/64).

[34 ]John Stuart Mill, 120. Cf. Mill to Gomperz: “My book on Hamilton is now finished, with the exception of a final revision which I shall give it a few months hence before sending it to press” (LL, CW, XV, 954; 22/8/64).

[35 ]LL, CW, XV, 963 (28/10/64). He here is looking forward to De Morgan’s paper on Infinity, because, as he says, the topic is touched on in the Examination (where De Morgan’s paper is not mentioned).

[36 ]Longman Chronological Register, 1860-77, f. 56, in the Longman Archive, University of Reading. On 11 March, he told Herbert Spencer he would soon offer him a copy of the work which (he vainly hoped) would contain “little or nothing to qualify the expression of the very high value I attach to your philosophical labours” (LL, CW, XVI, 1011).

[37 ]LL, CW, XVI, 1041 (to Longman, 30/4/65).

[38 ]Longman Chronological Register, 1860-77, f. 60. (Longman Impression Book 15, f. 158, gives August, but the Division Book for the period, also in the Longman Archive, confirms the July dating.) See also LL, CW, XVI, 1090-1 (to Spencer, 12/8/65), which concerns the note dealing with Spencer’s repudiation of part of Mill’s account of his views, the note being appended to the second edition of the Examination (Mill saw Spencer’s review too late for other treatment), and then placed where it belongs in the third edition (see 143 below).

[39 ]Autobiography, 163-4. Not all of Hamilton’s students, it may be noted, were unequivocally opposed to Mill’s views, for Fraser and Masson, as Mill indicates in his Preface, were not in agreement with most of the Edinburgh alumni.

[40 ]See LL, CW, XVI, 1223 (25/12/66), which refers to Mill’s desire to identify Grote as the author of the Westminster article on the Examination, and also to Bolton’s Inquisitio Philosophica. And cf. ibid., 1068 (18/6/65).

[41 ]John Stuart Mill, 124.

[42 ]LL, CW, XVI, 1161 (to Longman, 28/4/66); only 150 copies were then still in hand (Longman to Mill, 25/4/66; British Library of Political and Economic Science, Mill-Taylor Collection, I, #96, f. 226). By June the stock was down to 117 copies (Division Book).

[43 ]Longman to Mill (30/4/66), Mill-Taylor Collection, I, #95, ff. 223-4.

[44 ]LL, CW, XVI, 1223 (to Grote, 25/12/66).

[45 ]On 26 May he promised to send J. E. Cairnes a copy on publication (ibid., 1271), but Longman Chronological Register, f. 79, gives 16 May as the date of publication, and is supported in the May dating by the Division Book.

[46 ]Ibid., 1238 (9/2/67), and 1239 (11/2/67).

[47 ]Ibid., XVII, 1879 (6/4/72), and Longman Impression Book 18, f. 238.

[48 ]John Stuart Mill, 128.

[49 ]The format and type sizes remained constant through the editions, which were all printed by the same firm, Savill and Edwards (Savill, Edwards and Co. for the third and fourth editions). The principals in the firm of Longmans were, as might be expected, playing managerial chairs during these years: the first edition (1865) was published by Longman, Green, Longman, Roberts, and Green; the second edition (a few months later in 1865), by Longmans, Green, and Co.; the third—and, surprisingly, the fourth—(1867 and 1872) by Longmans, Green, Reader, and Dyer.

[50 ]Actually, the note concerning Spencer, mentioned above, appears on page 561 of the second edition, but became, as Mill intended, an incorporated footnote in subsequent editions.

[51 ]The count is also misleading because each of the editions was totally reset and, even though the type sizes were maintained, the proportions of text and footnotes were in some cases altered, resulting, with some changes to and from long and short pages, in different amounts of blank space at the ends of chapters. And, of course, with the additions, those blank spaces varied in size from edition to edition.

[52 ]See also xci-xciii below.

[53 ]These essays are in Essays on Philosophy and the Classics, Collected Works, XI (Toronto: University of Toronto Press, 1978), 245-69, and 449-71.

[54 ]One may also infer that this double consideration kept the matter in Mill’s mind when one looks at “Berkeley’s Life and Writings,” 455 and n.

[55 ]Here one sees another example of a softened tone: in the first edition the sentence began: “Mr. Mansel, entirely missing the point of this argument . . .” (see 85t, and cf. u-u).

[56 ]LL, CW, XVI, 1108 (25/10/65). The method (described below) of recording variants in this edition meets Mill’s criteria, but one dare not infer that he would therefore approve of it.

[57 ]In this edition the passages appear at 45n, 72n-4n, 216n-17n, 463n, where they are signalled (as are all variants) by superscript letters keyed to variant notes on the page. Much as one would have liked to adopt Mill’s method, it is, as the account above will hint, and any attempt to apply it will show, woefully inadequate.

[58 ]See 72n-4n, 74n, 93n-4n, 107n, 421n.

[59 ]This passage is even more confusing, for Mill adds a mention (without title or page) to Mansel’s “rejoinder,” while retaining the original reference (with title and page), which appears to apply to the (unsignalled) new citation.

[60 ]It could be argued that the added part of the note is a type (3) change; as will be evident, the categorization is more useful than certain.

[61 ]The change in the text would be treated as type (3) (a qualification) were it not for the explanation in the note.

[62 ]Calderwood was given a LL.D. by Glasgow in 1865 (subsequent to Mill’s first edition). The case of Ward is somewhat more puzzling, for the honorary degree of Doctor of Philosophy had been conferred on him by Pope Pius IX in 1854, long before the first edition of the Examination. How or when Mill became aware of the dignity is not known (he addressed Ward in letters simply as “Dear Sir”), and we are forced to fall back on R. H. Hutton’s amusing animadversions on the matter: “ ‘Ideal’ Ward was his Oxford nickname; ‘Squire Ward’ was his title in the Isle of Wight, where he had estates; ‘Dr. Ward’ was the description by which he was best known to the Catholic theologians; while his friends knew him simply as Mr. Ward.” Later, Hutton says of the Metaphysical Society (founded in 1869), “There, indeed, he was ‘Dr.’ Ward . . .” (quoted in Wilfrid Ward, William George Ward and the Catholic Revival [London and New York: Macmillan, 1893], 375 and 378). But Mill refused to join the Metaphysical Society, and one cannot suggest that he ceased to regard Ward as a friend.

[63 ]Changes of the contrary kind occur in other of Mill’s works, for example in the Logic (CW, VIII, 649e-e) where “Mr. Mill” became “Mr. James Mill” in 1868, presumably to make what had become a necessary distinction; were it not for these changes elsewhere, one might assume that the revision in the Examination was simply a third edition correction of the erroneously maintained courtesy (a Freudian slip?) to the living in the first edition.

[64 ]Perhaps it was no longer needed, because Bain—though still at Aberdeen—had gained, in Mill’s view, national (at least) celebrity.

[65 ]The addition is to part of a note added in 1867, which could just as easily at that date have concluded with the reference. The 1867 change is in one of those additions marked off by square brackets in a footnote; here again the diligent reader of the fourth edition will be confused, for the brackets were retained in 1872, with the further addition placed inside them.

[66 ]It is just possible—illustrating the problems of a nearly impressionistic classification—that we have here a type (2) change: was there some discussion (unknown to us) of the reliability of Pasteur’s work that came to Mill’s attention in the brief period between the first and second editions?

[67 ]This is seen as a type (3) rather than a type (1) variant because it does not evidently indicate either a change of judgment or new information; it will be evident, of course, that a very indistinct line divides the two classes.

[68 ]Since Mill is reporting others’ views, the change is prima facie less significant than the variant in the People’s Edition of On Liberty where the shift from species to individual has a much more marked effect on the sense. See Essays on Politics and Society, CW, XVIII, 224.37, and XIX, 657.

[69 ]An explanation of the decisions made here will be found in the headnote to Appendix B.

[70 ]Mill’s square brackets indicating (though he does not explain that they do so) tampering with Hamilton’s text were placed correctly in the first and second editions: i.e., the text read as it is here printed, “that I [believe]” and “conscious of [the”. In the third edition the reading (retained in the fourth) was “that [I believe]” and “conscious [of the”.

[71 ]In the first and second editions (the reading here accepted), Archbishop Whately is noted to have said that Logic is both a Science and an Art “in an intelligible sense.” In the third and fourth editions, in what at first glance is acceptable, the sentence ends, “is an intelligible sense.”

[72 ]The reading in the first three editions (the final colon introduces a quotation from Hamilton) is: “Sir W. Hamilton, therefore, needs another kind of argument to establish the doctrine that the Laws of Identity, Contradiction, and Excluded Middle, are laws of all existence: and here we have it:” but in the fourth edition “leave” replaces “have” in the last clause. Here again at first glance the final reading is plausible (partly because of the uncertain reference of “it,” which, if taken to refer to “doctrine,” would make the final reading more intelligible, but which, if—as seems right—taken to refer to “argument” makes the original reading more likely), but is rejected in our text.

[73 ]Only slightly less plausible than the previous two examples is the mistaken substitution here in 1867 of “qualities” for “quantities.”

[74 ]The omission or misplacing of quotation marks in the passage from Hamilton concerning Descartes makes it extremely difficult to determine who is being quoted; the difficulty becomes an impossibility because of Hamilton’s errors in translation and ascription, until his sources are consulted.

[75 ]The reason for listing these as variants is consideration for the reader, hypothesized by Mill in a passage quoted above, who may own an early edition that he wishes to compare, and might be surprised by the differences in these two places. In the first, “unnecessarily” appears rather than the correct “necessarily”; in the second, five lines (in the original setting) were omitted (both corrections were made in 1867).

[76 ]The first of these is the change, in the second edition, from “coals” to “coal” in the clause “the water, and the wood or coal, were not destroyed.” Mill may have carelessly used the plural form (in British usage, indicating individual bits of coal), or, of course, the compositor may have misread his hand (cf. the next footnote) or simply made an error. The second, which occurs within a quotation, looks like a conscious change by Mill, in the second edition, in an attempt (unsuccessful by grammatical standards) to improve Hamilton’s sense by changing “what does this imply?” to “what does it imply?”

[77 ]It may be noted that overall only seven changes reverse changes in earlier editions.

[78 ]The editions of the works on logic which Mill cites by Aldrich, Bartholinus, Brerewood, Du Hamel, Fell, Keckermann, Smith, and Wallis, at present in the London Library, were very probably given by Mill to the Library with other of his father’s books. The last two are autographed “J. Mill” on the title page. Of the others, Ammonius, Burgersdijck, Du Trieu, and Sanderson are (or were) in his library, Somerville College.

[79 ]In many of his review essays, however, quotation bulks very large indeed, because the great periodicals saw, as one of their functions, making reviewed works known by their contents to their subscribers; in the Examination Mill’s intention was to expose rather than reveal.

[80 ]LL, CW, XVII, 1613 (7/6/69).

[81 ]Veitch, Memoir, 196n-7n, quoting Examination, 492-3 and 438-9.

[82 ]Veitch, 381-3, quoting Examination, 499.

[83 ]Veitch, 429-48, referring to Examination, 498n-9n, 499-500 and 501-21, and 503.

[84 ]Veitch, 404 and 404n (the whole discussion occupies 404-20).

[85 ]He and Calderwood (who was second on Hamilton’s prize list in 1847) were classmates in Hamilton’s logic class of 1850.

[86 ]Veitch, 379 and 435. Mill also compares the Lectures to Hamilton’s “later speculations” (372); in objecting to Mill’s use of the Lectures, Veitch bears out this comparison, pointing out that they were prior to “nearly all” the footnotes and all the “Dissertations on Reid”—indeed to everything but the early Edinburgh articles in the Discussions (209-10).

[87 ]Veitch, 446.

[88 ]Ibid., 421-8. Since Mill chose not to notice so much, it may be mentioned that Veitch chose not to mention one instance of Mill’s expressing a judgment that Veitch would have appeared to accept. In discussing Hamilton’s odd and unoriginal treatment of mathematics, Mill mentions that Hamilton is much inferior to Dugald Stewart on the matter, and that the “cloud of witnesses” Hamilton summons makes us no wiser (470). Veitch does not quote Mill, but includes the following passage from a letter to Hamilton from Napier about Hamilton’s article on its first appearance: “One criticism [of the article, by various people], I confess, I was not quite prepared for—viz., that the argument is injured by the ‘cloud of witnesses,’ which, it is said, has been huddled together without discrimination, and without any rational view of the value of authorities. Lord Brougham, in a letter I received yesterday among others, makes this remark, and adds, that he is sorry the writer, whom he praises for ability and learning, should have adopted a tone in regard to mathematics so different from that of the cautious and philosophical D. Stewart.” (174; 1/2/36.) Lest it be thought that Brougham and Napier are quoting Mill thirty years in advance, it should be noted that Hamilton himself refers in his article to the “cloud of witnesses”—and, indeed, they are all quoting St. Paul (Hebrews, 12:1).

[89 ]Moral Causation, 2nd ed., iv. Alexander’s works may be recommended to connoisseurs of what—were it not for Sir William Hamilton—might be thought of as the Edinburgh style.

[90 ]It may be noted that McCosh calls forth some of Mill’s most acerb remarks, among them this: “I must add, that the chapter of Dr M‘Cosh from which I am now quoting, that headed ‘The Logical Notion,’ contains much sound philosophy, and little with which I disagree except the persistent illusion which the author keeps up throughout the chapter that I do disagree with him” (317n). Hamilton does not escape this kind of comment, of course: see, e.g., “Sir W. Hamilton . . . thus, as is often the case (and it is one of the best things he does), saved his opponents the trouble of answering his friends . . .” (445).

[91 ]Cambridge: Deighton, Bell; London: Bell and Daldy, 1865.

[92 ]Mansel, of course, is dealt with frequently in other places, most extensively in the first quarter of the work, but also in Chapters xvi and xvii. It may be noted also that xiv, on Association, was particularly enlarged in 1872, perhaps because Mill had edited his father’s Analysis in the preceding years.

[93 ]It should be mentioned that the majority of these additions come in the appended notes to xi and xii (combined), xiii, and xiv.

[94 ]Mill’s consummate expository skill, his level tone, and the speed with which he enters into topics all disguise what are sometimes rather weak transitions in the Examination. Much detail would be here inappropriate, but one may note briefly that, while Mill moves generally from metaphysics to logic (thereby, as a glance at his footnotes will show, following Hamilton’s Lectures), with the final five chapters dealing with special topics, the flow is not always smooth. The first six chapters, constituting the main assault on Hamilton’s metaphysics, hold together well enough, but Chapter vii, on Mansel (whose ideas admittedly are involved from iii-vi), is interpolated. There is no lead from it to viii, and the opening of viii (which suggests Mill’s uneasiness over the arrangement) refers back to the matter of ii ff. and vi, and hints at what will come later. From viii through x we have exposition and criticism of some of Hamilton’s psychological views; x is paired with xi, which gives Mill’s views, the empirical position being continued through xiii, by which time one is quite far from Hamilton, to whom (and to Mansel) the argument returns in xiv. No persuasive transitions occur between xiv and xv, xv and xvi (in xvi it would have been appropriate to refer back to viii, where this discussion is promised), or xvi and xvii. Logic is the unifying thread from xvii through xxiii, and the little structuring help Mill gives is adequate. There is an acknowledged leap from xxiii to xxiv, as Mill indicates his move from matters purely psychological and logical to questions relating to the Philosophia Prima (or better, Ultima), but these are all dealt with in the one chapter, one must suppose, for xxv reveals a clear signpost, though it is one that suggests the path behind has been meandering: we are now done with the main part of Hamilton’s psychology, that on Cognitive Faculties, leaving Feelings and Conative Faculties. But Hamilton, Mill says, barely treats of these, and the following discussion in xxv is solely on pleasure and pain. Only a weaktransition to xxvi, on the freedom of the will, occurs, and xxvii, on Hamilton’s view of mathematics, is justified solely on the ground that the examination would be incomplete without it. The concluding chapter does its work well, giving—as is typical of Mill—a little ground, only to seize it back again.

[95 ]One special departure from the normal practice deserves mention. At 399n.20 “All A is B” is corrected to “All A is all B” even though the four editions have the former reading. The justification for this emendation is that the sense is lost without it, and that elsewhere in the passage, as well as in the authors referred to, the latter reading is always found in analogous contexts.