### Peter H. Nidditch (ed.), The Clarendon Edition of the Works of John Locke: An Essay Concerning Human Understanding

Find Location in text

# A. GENERAL

## (1) The Significance of the Essay

Locke's Essay Concerning Human Understanding is the primary classic of systematic empiricism. Locke's empiricism is realistic and rationalistic: he demands the rejection of scepticism, and of the pursuit of verbal issues and wrangling; for him, 'Reason must be our last Judge and Guide in every Thing';1 and he insists that the ultimate source of all our ideas and the ultimate required test of all our putative knowledge and beliefs lie within the bounds of the workings of normal sense- or inner-experience, whose modest potentialities should be methodically elucidated and their fulfilment circumspectly approached. The contents (apart, perhaps, from the fundamental theism) of the Essay were a philosophical reflection, cultivation, and further promotion of attitudes and assumptions characteristic of critical and cognitive thought in an Age of Science. The Essay's success in winning the attention of the European public from the end of the seventeenth century and through the eighteenth century and after by its account of the origin, modes, and scope of human understanding was similar to the triumph of Newton—the outcome of his Principia and Optics—concerning the physical world. The message of the Essay became deeply diffused among successive generations of philosophers, and of leaders of educated opinion, in Britain and abroad. Far outmatching any of its forerunners (for example, Bacon's Novum Organum) in its systematic elaboration and in its impact, Locke's Essay is the vital ancestor of all later empiricism, and of Psychology.

The Essay is thus among the most significant works in the history of thought; it ought, then, to be the subject of an edition that accords with the canons of modern textual scholarship, this being all the more necessary because there is abundant evidence of Locke's preparations and revisions of it over more than half his adult pg xlifetime. (No philosophical classic by Bacon, Hobbes, Descartes, Spinoza, Leibniz, Berkeley, Hume (his Essays excepted), Kant, Hegel, or anyone else in their period is comparable to the Essay in respect of the frequency of authorized and of considerably amended editions of it. Montaigne's Essais, Mill's A System of Logic, Ratiocinative and Inductive, and Mach's Die Mechanik in ihrer Entwicklung, historisch-kritisch dargestellt are among the few philosophical classics outside that period that are comparable to it in this respect—but, then, not in respect of the quantity or extended temporal distribution of extant manuscript drafts for the book.)

## (2) The Objectives of this Edition

The present critical-text edition of Locke's Essay offers a text that is directly derived, without modernization, from the early published versions; it notes the provenance of all its adopted readings (some of which are new, correcting long-established errors); and it aims at recording all relevant differences between these versions. These differences are numerous, requiring several thousand registrations both in the case of material variants (deletions, additions, or changes of wording) and in the case of formal variants (changes of punctuation, parentheses, italics, etc.). (Many registrations cover more than one alteration each.) The main critical apparatus, recording the material variants and the principal semantically significant variants of punctuation, accompanies the text at the foot of the page; a separate register covering the other variants is placed on its own at the end. Each page of text carries line numbers in order to facilitate exact study and reference.

The Fourth Edition of 1700 has been chosen as copy-text; it is demonstrable that in formal respects this Edition2 is closer to Locke's wishes so far as these are known than the posthumous Fifth Edition of 1706, although certainly the material alterations which first appeared in the latter must be appropriately accommodated. Besides these two Editions, full account has also been taken pg xiof their predecessors, and of Coste's French translation of 1700; in addition, the later English publications of the Essay in 1710 (the sixth edition), 1714 (the first edition of Locke's collected works), and others till 1730, and Coste's revised translation of 1729, have been examined.

No previous scholarly edition of the Essay has been serviceable as a basis for the present volume. Fraser's edition (Oxford, 1894) is pervasively faulty; it gives a modernized text, which is repeatedly inaccurate; it records among its annotations only a fraction (equal to about five per cent) of the material variants—and then frequently in a confused way; and it notes a negligible number of the formal variants.3 Yolton's edition (London, 1961), whose text is a modernization of the Fifth Edition, contains in total many errors or dubious pg xiireadings,4 and it purposely has no critical apparatus. No modern edition presents a complete and reliable text that faithfully preserves the original in all its essentials, or provides a comprehensive record of the alterations of Locke's thought and its expression in the printed Essay.

I hope that my text and critical apparatus will be found useful and reliable by philosophers and by historians of thought and literature, as well as by Locke specialists. In view of the new editorial matter and of the great number of words, letters, and pointings in the various versions and the present text of the Essay concerning which oversights or mistaken judgements could have occurred here, one can hardly expect an entirely impeccable result on this first appearance of a critical-text edition of the Essay. I shall naturally be grateful to be informed of any required corrections.

It is intended to supplement this edition in due course by a further publication (henceforth referred to as Drafts: short for Drafts for An Essay Concerning Human Understanding), towards whose preparation some progress has already been made; it will contain transcriptions of Lockian manuscript material relating to the history of the Essay, largely but not exclusively prior to 1690; the transcriptions will be full ones, incorporating and marking all cancellations, substitutions, and additions.5 I plan to survey as a whole the context and evolution of the Essay in the Introduction to Drafts. (See also pp. xxxvii f. and n. 90 below.) The Introduction to the present edition proceeds to concentrate on summarily providing a description of the making and make-up of the early versions and an account of the method I have used to edit and compare their contents.

# B. THE EARLY VERSIONS OF THE ESSAY

## (1) The First Edition, 16906

Locke's thinking and writing about the Understanding that eventuated in the published Essay of 1690 began approximately pg xiiitwenty years earlier.7 Two manuscripts that were among his original efforts in this matter carry the date 1671. One of these manuscripts (Draft A) is headed 'Sic Cogitavit de Intellectu humano Jo: Locke an̄ 1671 / Intellectus humanus cum cognitionis certitudine, et assensûs firmitate';8 the other manuscript (Draft B), which is much longer and presumably later, has these titles on consecutive pages: 'Intellectus / 1671 / J. L.'; 'De Intellectu humano / 1671 / An Essay'; and 'An Essay concerning The Understanding, / Knowledge, Opinion & Assent'.9 The first appearance, in an extant draft, of the title of the published Essay was fourteen years later: 'An Essay / Concerning humane understanding / in fower books / 1685'.10 However, Locke continued to refer to his work sometimes under the name 'Intellectus' or 'De Intellectu' in his correspondence and elsewhere.11

It appears, from a letter of his to Clarke, 1 January 1685 (N.S.), that Locke commenced in 1683–4 the rewriting of the De Intellectu that led, without a break but after further revisions, to the published Essay. The first complete draft of the four Books of the Essay as we know it was not finished until December 1686; Locke acknowledged pg xivthat it contained too many repetitions and was in need of some re-ordering.12 He proceeded to revise it in the course of the next two and a half years. At the same time, he prepared an Epitome of the Essay; this was translated into French by Jean Le Clerc and published early in 1688 in Le Clerc's Journal Bibliotheque universelle et historique.13 It is likely that two or more copies were made of the 1685–6 draft (or unfinalized revisions of this draft) of the Essay, Draft C being a portion of one such copy. (More than one copy existed of a pre-1685 draft—possibly Draft B or possibly a draft made in the early 1680s—for Tyrrell possessed and long kept such a copy from late in 1683 onwards, as is shown by his letters to Locke, (c. 23) February 1683/4, 29 August 1687, and (c. 19) December 1689; cf. also Locke's letter to Clarke, 1 January 1685 (n.s.).) When the Earl of Pembroke wished to see the whole work and clearly hinted that he would be willing for it to be dedicated to him (a procedure he demurred at in respect of the Epitome),14 Locke asked Clarke to forward the copy which he had to Pembroke;15 and Locke must have kept at least one copy for his own needs. Locke's final revised draft for publication probably existed in two (not necessarily 'fair') copies, one of which would have become used as printers' copy, while Locke retained the other. No copy of any part of any pg xvdraft of the Essay dating from 1685 to 1689 has survived with the exception of Draft C. In particular, the printers' copy met the usual destructive fate of manuscripts serving this purpose in the sixteenth, seventeenth, and eighteenth centuries (but see n. 66 below).

Locke entered into a contract, dated 24 May 1689, with the London bookseller Thomas Bassett to publish the Essay.16 It was agreed that the printing would be begun immediately and be continued at the rate of at least four sheets a week for a volume in folio; and, further, that Locke would deliver to Bassett, on demand, material equivalent to two such sheets upon Bassett's delivery to him of a copy in print of what Bassett had last received from him. There was a total of about 100 sheets. The contract provided for Locke to be paid at the rate often shillings per 'sheet', defining a sheet for this financial purpose as being the equivalent of so much as was contained in a sheet of Milton's History of England in octavo, within seven days after the printing of the book had been finished; there were 58 'sheets'. Bound copies of the book were received by Locke on 3 December 1689 and copies were on sale within a week or two in London and Oxford.17

It has been suggested—without, however, an explanation being proposed—that if printing began, and proceeded, as agreed to in the contract, 'one might have expected the 58 agreed sheets to have been finished in fourteen and a half weeks, by the beginning of September, and not the beginning of December'.18 Evidently, however, we must not confuse the number of sheets in Locke's book with the notional, only financially relevant, number of 'sheets' in it, we must allow time for the bookbinder to bind the copies Locke received from him on 3 December, and we must take account of other available evidence. In a letter Locke wrote to Limborch, 3 December 1689, he expressed the hope that the last sheet of his treatise would be printed off that day ('Hodie ultimam tractatus mei de Intellectu humano schedam typis impressam esse spero').19 pg xvi(An entry in the Errata corrects a reading on the final page of the main text; doubtless the matter printed last was the title-page, the preliminary Epistles and page of Errata, and the Contents pages at the end.) Quite possibly the delay of a week or two in completing the printing of the book was partly the fault of the printer—but the compositors' and pressmen's work did proceed at the high rate of practically four sheets a week.20 It was commonplace in England in the seventeenth century for authors to bait printers for their unreliability in one respect or another. Locke shared in this practice of complaint, especially in the case of his Two Treatises of Government. In the letter to Limborch quoted above, Locke immediately goes on to remark 'so at least the printers have promised, if any reliance is to be placed on the word of these men'. But Locke himself was perhaps partly responsible for the slight delay. Firstly, the contract suggests that he did not let the printers have at the outset a complete manuscript of the Essay, copy being passed on to them in batches; I suppose this was because he wanted to be able to make revisions up to the last minute.21 This procedure hardly promoted a speedy execution of the printing work; time could easily have been lost by a messenger's having to carry printed sheets to Locke at his lodgings in Dorset Court in Westminster, possibly wait there to collect required new copy, and then travel back to the printing-house with it.22 Secondly, Locke may have wished to make changes after he received printers' proofs; the terms of the contract with Bassett make it clear that Locke intended to check each part of the Essay after it was printed (cf. p. xxiv below). The printing-house could not, owing to its limited supply of the founts and other pg xviiequipment being used, keep more than a few sheets of the book in type before printing them off and distributing the type for further composition; hence, once such a distribution had been made, an Errata page was required for embodying corrections—unless each affected sheet was recomposed as far as necessary to accommodate alterations. (The different contributions of the printing-house proof-reader, and of Locke and his friends, to the Errata page in the First Edition is unknown; many printing errors are not listed in the Errata.)

There are two issues of the First Edition. Apart from some typographical details, the title-page of copies of one issue runs: 'AN / ESSAY / CONCERNING / Humane Understanding. [rule] In Four BOOKS, [rule] [Epigraph from Cicero] [double rule] [wreath] [double rule] LONDON: / Printed by Eliz. Holt, for Thomas Bassett, at the / George in Fleetstreet, near St. Dunstan's / Church. MDf''pos; The title-page of copies of the other issue is similar except for the imprint at the foot, which runs: 'LONDON: / Printed for Tho. Bassett, and sold by Edw. Mory / at the Sign of the Three Bibles in St. Paul's / Church-Yard. MDCXC.'

For my first memoir on the Essay23 I had presumed, along with the majority of those acquainted with the two issues, that the Holt issue was prior.24 My reasons then were (1) that only the Holt issue was mentioned by Locke in his notebooks, etc.,25 (2) that only the Holt issue was listed in the booksellers' Term Catalogues,26 (3) that Bassett was the only bookseller with whom Locke had agreed terms for the publication of the Essay, and (4) that the title-page of the Mory issue was a cancellans. The question of priority has since been re-examined by others (see nn. 18 and 24)27 and the majority view that had been held has been ratified.

pg xviiiIn contrast with his books on Toleration, Government, Christianity, and Education, none of whose First Editions has his name at all, Locke at once acknowledged his authorship of the Essay by appending his name to The Epistle Dedicatory that it included, although his name was absent from the title-page in the First Edition.

The question how many copies were printed of the First Edition of the Essay seems, like the proof-reading aspect, not to have been previously discussed by Lockian scholars; the number of copies I shall suggest, and similarly the numbers of copies to be suggested below of later Editions, must be regarded as entirely provisional. (a) It has often been claimed that there were only a few hundred copies printed of the first edition of Newton's Principia (1687). A book as advanced and technical as the Principia could naturally have been expected, in the planning of a first edition, to have a smaller reading public than a discursive, philosophical book such as the Essay; so the minimum number of copies of the Essay is, say, 250.28 If one knows or can reasonably surmise the trade or retail price, a better way of determining this minimum might be to calculate the economic minimum—for the publisher to avoid a loss. (b) At the other extreme, the normal maximum number of copies in an edition of a book in England in this period was 1,500. There are five other relevant sorts of evidence: (c) the well-established bookseller Awnsham and John Churchill produced, in suitable cases, a little more than 1,000 copies of a book;29 (d) Bassett was chiefly a law publisher (but he did have a varied list, e.g. he shared in the publishing of the first edition (1686) of Sir Thomas Browne's collected works); (e) he in part sub-contracted with Mory in the publication of the Essay; (f) the time needed for selling out the First Edition was about three and a quarter years;30 (g) the number of copies that have survived. I estimate that about 900 copies of the pg xixFirst Edition were printed, by far the greater number of them belonging to the Holt issue.

The First Edition was published in the form of a folio volume, with the following contents after the title-page: The Epistle Dedicatory (4 unnumbered pages); The Epistle to the Reader (5 unnumbered pages); Errata (unnumbered page; 120 entries); Bk. I, Chs. i–iv (pp. 1–36); Bk. II, Chs. i–xxxi (pp. 37–183); Bk. III, Chs. i–xi (pp. 185–260); Bk. IV, Chs. i–xx (pp. 261–362); The Contents [headings of books and chapters] (2 unnumbered pages); The Contents [headings of chapters and summaries of contents of chapter sections] (20 unnumbered pages). There are three mispaginations: 287 was numbered 269; 296, 294; and 303, 230. Generally, the pages of the main text contained 54 lines, with about 13 words per line; each page had a running (normally, chapter) title, and a catchword at the end. In this Edition, as in the later Editions in folio, the signatures are arranged for gatherings to contain two sheets, i.e. four leaves (eight pages). The volume had no portrait, marginal sectional summaries, or Index.

## (2) The Second Edition, 1694

(i) In his letter of 20 September 1692, Locke asked Molyneux 'whether it would not be better now to pare off, in a second edition, a great part of that which cannot but appear superfluous to an intelligent and attentive reader. If you are of that mind, I shall beg the favour of you to mark to me these passages which you would think fittest to be left out.' Molyneux, expressing his own opinion and that of 'two very ingenious persons here' to whom he had lent a copy of the Essay, assured Locke that omissions were not required. 'I never quarrelled with a book for being too prolix, especially where the prolixity is pleasant, and tends to the illustration of the matter in hand, as I am sure yours always does ... the work in all its parts [is] so wonderfully curious and instructive.'32 This verdict was decisive for Locke's treatment of the Essay for the purpose of revised editions and therefore for the historical effect of the Essay: he did not attempt to omit or abridge any chapter at all, and his modifications of the book made it, on the contrary, increasingly longer. If Molyneux had firmly recommended certain moderate measures of reorganization and abbreviation of the book's contents, Locke would undoubtedly have been willing to make the necessary changes, for he admitted, with regret, the existence of repetitiousness and was seeking, in Molyneux, a helping hand towards the reduction or elimination of this. (Cf. n. 50 below, and 8(2)–9(11).)

(ii) Changes in Bk. II, Ch. xxi, 'Of Power'. A large part of the chapter was new in the Second Edition. In his letter of 22 December 1692, Molyneux remarked that Locke's 'Discourse about man's liberty and necessity' required 'some further explication'. Locke outlined his new 'hypothesis' in a letter to Molyneux, 15 July 1693, pg xxiwhich contained the proposed summaries of §§ 28–40 of this chapter.33

(iii) The new chapter 'Of Identity and Diversity', inserted in Bk. II as Ch. xxvii (so that the original chapter thus numbered and the following chapters of this Book were renumbered in the Second Edition). 'You will herewith receive a new Chapter Of Identity and Diversity, which, having writ only at your instance, 'tis fit you should see and judge of before it goes to the press. Pray send me your opinion of every part of it. You need not send back the papers, but your remarks on the papers you shall think fit, for I have a copy here.'34

(iv) Molyneux's Problem (whether a person born blind, and able in that state to distinguish by touch between a spherical and a cubic solid 'nighly of the same bigness', would, if made to see, be able to distinguish them by sight).35 A statement and discussion of this problem was added to Bk. II, Ch. ix, § 8.

(v) Molyneux expressed doubts about a number of particular passages, because of their apparent inaccuracy, unclearness, or inconsistency; § 22 of the new chapter 'Of Identity and Diversity' in draft was the most persistently discussed of these passages.36

(vi) The 'Errata Typographica (besides those mentioned in the pg xxiiTable) are many and great; these therefore, in your next edition, are diligently to be corrected', Molyneux wrote to Locke, 22 December 1692; and three months later Locke acknowledged Molyneux's help in making such corrections: 'The faults of the press are, I find upon a sedate reading over of my book, infinitely more than I could have thought; those that you have observed, I have corrected, and return you my thanks.'37

(vii) Locke agreed to Molyneux's proposals that the sectional summaries should be added in the margin of the text, and that there should also be an Index.38

Locke was, however, unwilling to make some other alterations suggested by Molyneux; for example, the possibility of insisting 'more particularly, and at large, on Æternæ Veritates' was mentioned by Molyneux, to whom Locke replied about this in due course by saying: 'You desired me too to enlarge more particularly about eternal verities, which, to obey you, I set about, but upon examination find all general truths are eternal verities, and so there is no entering into particulars; though, by mistake, some men have selected some as if they alone were eternal verities.'39

The Second Edition contained only one chapter ('Of Identity and Diversity') that was wholly new, and only one chapter ('Of Power') that was extensively reworked and given many new additional sections. An analytical Index was also now added ('I intend to make an alphabetical index to the book', Locke wrote to Clarke, 19 March 1693/4). Besides these additions, Locke made numerous, more limited ones throughout the Essay. The whole book was carefully examined with a view to making material or formal improvements, and wherever it seemed desirable it was amended. (Separate printed slips containing all the longer alterations and additions were produced for owners of the book in its First Edition to incorporate in their copy. Cf. 12(19), n.) Apart from the many cases where one word was replaced by another, the majority of the large number of other alterations of wording consisted in adding one or more words, or in replacing one or more words by a longer group of words. Locke's revision was not restricted to matters of wording; he was an author who concerned himself also with the pg xxiiiformal aspect of his books.40 The First Edition of the Essay was altered in italicization, the use of parentheses, pointing, and in other formal respects, in a large number of places; it is a reasonable surmise that Locke gave the instructions for the alterations of the first two types, for at least almost all the alterations involving punctuation signs heavier than the comma (e.g. a change from a semicolon to a full stop or vice versa), and for a proportion of the other alterations (especially in the use of the comma).41

From early March 1688 (n.s.), Locke went to a lot of trouble to regain possession of a portrait of himself which John Greenhill had painted in about 1672 and which he wished to have reproduced as an engraved frontispiece in copies of the Essay. According to Locke, he had lent this portrait to Mr. and Mrs. Stringer for safe-keeping during his absence abroad; the Stringers insisted that Locke had given them the painting as a gift. Eventually, in the summer of 1689, the portrait was made available to Locke to enable him to have an engraving made from it; but, for whatever reasons, Locke's original plans for a frontispiece fell through.42 For the Second Edition, an engraving was made from a portrait of Locke drawn by his servant and amanuensis Sylvanus Brounover (probably using the Vereist portrait of Locke, 1689); this engraving was used for a frontispiece included in copies of the new edition. Brounover's portrait continued to serve as the source of the frontispiece in the subsequent Editions of the Essay.

Locke began actively in 1692, if not earlier, to revise the Essay for republication; and he made alterations up to the last minute—in the Errata of the new version—until the Second Edition issued from pg xxivthe press with his name now on the title-page, in May 1694.43 The commencement of the type-setting for the Second Edition was in November or December 1693: in his letter to Molyneux, 23 August 1693, Locke wrote 'My Essay is now very near ready for another edition' while in that of 19 January 1693/4 he was able to report 'my book ... advances now apace, and I believe there are, by this time, near 150 pages of it printed'; in his letter to Newton, 5 October 1693, Locke said with reference to the Essay 'My book is going to the press for a second edition'; and Locke's only visit to London in the relevant period was from early November till 19 December 1693, and its timing was surely in part intended for the purpose of his leaving printers' copy with one of his London friends and his making final arrangements concerning the printing of the new edition of the Essay. The production of the Second Edition was fairly quick; Locke's remark about 'the slowness of the press' in his letter to Molyneux, 26 May 1694, announcing the publication of the Second Edition, should be regarded as expressing an author's impatience to have his latest work made available and not interpreted as evidence of inefficiency by the printers.

Locke did not go to London again until 23 April 1694. He received and checked printers' proofs of the Second Edition at Oates in Essex where he was living; there is direct evidence of this from Manship's letter to him, 10 March 1693/4, which begins 'I have sent you what papers that are printed this week and shall continue sending until it is finished'. Locke followed a similar procedure for some of his other works: if he could not be in London to do it, he examined proofs at Oates. Thus, he wrote to Clarke from Oates, 2 November 1692, concerning Some Thoughts Concerning Education? 'I have this farther favour to beg of you, that you would send for Mr Awnsham Churchill (to whom I have writ four or five times to desire him to send me the sheets have been printed since I came out of town, but cannot receive a word from him) and tell him I would by no means have him publish it till I have perused all the remaining sheets, which I would have him send to me'; and other letters of his to Clarke (e.g. 7 March 1693/4, 12 March 1693/4, 30 March 1694) witness to his reading proofs for the second edition of Two Treatises of Government.

pg xxvAs mentioned above, Bassett had undertaken in 1693 to publish a new edition of the Essay. But he disposed of his rights in the book in late February or early March 1693/4 in equal shares to Samuel Manship and Thomas Dring; and the latter, in turn, promptly assigned his share to Awnsham and John Churchill. There were two issues of the Second Edition. A small number of copies, constituting presumably the first issue, carried an imprint in the joint names of Dring and Manship.44 Nearly all the copies, constituting the other issue (the only one listed in the Term Catalogues),45 carried an imprint in the joint names of the Churchills and Manship. The contents of the title-page of copies of this issue (and of later Editions) are indicated on pp. 1–2 below. I estimate that about 700 copies were printed of the Second Edition.

The Second Edition was published in the form of a folio volume, with the following contents after the title-page: The Epistle Dedicatory (4 unnumbered pages); The Epistle to the Reader (5 unnumbered pages); Errata (unnumbered page; 53 entries); The Contents [headings of books and chapters] (2 unnumbered pages); The Contents [headings of chapters and summaries of contents of chapter sections] (21 unnumbered pages); Bk. I, Chs. i–iv (pp. 1–39); Bk. II, Chs. i–xxxii (pp. 41–219); Bk. III, Chs. i–xi (pp. 221–99); Bk. IV, Chs. i–xx (pp. 301–407); Index (11 unnumbered pages). The format (except for the marginal sectional summaries) was the same as in the First Edition.

## (3) The Third Edition, 1695

The Second Edition immediately sold well. In a letter to Limborch, 26 October 1694, Locke told him 'Libri mei de Intellectu humano secunda editio distrahitur celerius quam credere possem'—'The second edition of my book on human understanding is selling more quickly than I could have believed.'46 Molyneux received a complimentary copy of the book in July 1694; in writing to him, 3 September 1694, Locke again invited his criticisms: 'Any faults you pg xxvishall meet with in reasoning, in perspicuity, in expression, or of the press, I desire you to take notice of, and send me word of. Especially if you have anywhere any doubt; for I am persuaded that, upon debate, you and I cannot be of two opinions.'47 Towards the end of that calendar year Molyneux dispatched a list of required corrections, which Locke acknowledged in a letter, 8 March 1694/5, adding 'I will take all the care of [them] I can in the next edition, which, my bookseller tells me, he thinks will be this summer' and inviting further corrections. (This evidence alone should have prevented scholars nowadays from commonly describing the Third Edition as 'a reprint', meaning by this a completely unrevised reproduction of the Second Edition.) 'The third edition of my Essay', Locke informed Molyneux, 26 April 1695, 'is already or will be speedily in the press'; by 2 July he was able to report, more firmly, that it 'is now in the press'.48 Locke was in London 30 May–22 June 1695, and it was possibly only then that arrangements for the printing of the Third Edition were finalized. Presumably he gave instructions that the Third Edition was to be an exact reproduction of the Second Edition except in the places for which he had noted amendments for the printers. The new edition was probably published in about December 1695. There is no separate evidence, in this case, of whether, and if so, where, he examined the proofs; he was in London 8 July–20 August, 2–28 September, and 21 October–18 November 1695 and could, in the course of these visits, have checked some passages. It is, however, not without significance that the Third Edition contained no page of Errata in spite of its host of misprints; I have no doubt that Locke did not check any proofs of it.

The Third Edition is closely similar to the Second in content and in format. The corrections listed as Errata in the Second Edition were incorporated in the reset text for the Third Edition; this also contained Locke's latest—only occasional and brief—material amendments (see e.g. 50(36), n.; 181(20), n.; 275(34), n.; 313(26–7), n.; 341(1), n.; 345(28), n.; 489(5), n; 656(13) and (14–15), nn.; 664(12), n.), and a small number of formal amendments in pointing and italicization that were presumably his (and that were retained in the Fourth Edition), e.g. at 60(27), 123(21), 338(29), 342(30). Other changes originated from the printers. Now and then a word (or a pair of words) in the Second Edition was erroneously replaced by another, or a word was erroneously omitted or inserted; none of pg xxviithese particular errors occurred in the Fourth Edition. Besides such material misprints, which were unfortunately frequent, the printers introduced far more numerous formal modifications, of a slight kind; these were most commonly changes of upper or lower case, of italic or roman type for colons, or of the apostrophe and 'e' in a past tense or a participle, and additions or omissions of commas: pointings heavier than a comma were generally reproduced with accuracy, only a small number being misprinted. In their formal aspect, most pages were identical with the corresponding portions of the Second Edition except for a few letters, apostrophes, commas, or changes of type for colons.

The pagination of the Third Edition for the most part coincides exactly with that of the Second Edition; thus, in both Editions, pp. 1–178 and 199–219 correspond, Book III begins on p. 221, Book IV begins on p. 301, and the last page of this last Book is numbered 407.

I estimate that 800 copies were printed of the Third Edition.

In January 1694/5, John Wynne (1666–1743), at that time a Fellow of Jesus College, Oxford, wrote to Locke offering to compile an abridgement of the Essay: 'I am inclined to think that it would be very useful to publish an abridgement of the book. If some of the larger explications (some of which are but incidental to the general design of the work) were contracted, it might be reduced to the compass of a moderate 8vo.'; Locke gladly gave his approval to this undertaking.49 It was published as a handy volume entitled An Abridgment of Mr. Locke's Essay concerning Humane Understanding, with 310 numbered pages, by the Churchills in March 1696 (N.S.); it carried a dedication to Locke dated 17 April 1695.50 Wynne's abridgement enjoyed a considerable success, and was often reprinted, until late in the eighteenth century. One of its most striking features was its disregard of the whole of Book I of the Essay other than the introductory first chapter.

## pg xxviii(4) The Fourth Edition, 1700

Even while the Third Edition was in the press, Locke was giving thought to further changes to be made to the Essay for later publication, including for the Latin translation that Molyneux was then trying to arrange. Thus, he mentioned to Molyneux, 8 March 1694/5, the idea of adding something on Enthusiasm in Bk. IV, Ch. xviii ('Of Faith and Reason, and their distinct Provinces') or in a new chapter on this topic; although this addition turned out eventually to be in the latter form (Bk. IV, Ch. xix in the Fourth Edition), he seemed in his next letter to Molyneux, 26 April 1695, to have decided on the former alternative, remarking that he designed this addition for Bk. IV, Ch. xviii 'as a false principle of reasoning often made use of'. At the end of this letter he made reference to what became another new chapter (Bk. II, Ch. xxxiii, 'Of the Association of Ideas'): 'I think I shall make some other additions to be put into your Latin translation, and particularly concerning the Connexion of Ideas, which has not, that I know, been hitherto considered, and has, I guess, a greater influence upon our minds than is usually taken notice of.'

Those two chapters were the only new ones in the versions of the Essay that were published after the Second Edition. In 1697, Locke had plans for another new chapter; once again he confided his incipient thoughts to Molyneux: 'I have lately got a little leisure to think of some additions to my book, against the next edition, and within these few days have fallen upon a subject, that I know not how far it will lead me. I have written several pages on it, but the matter, the farther I go, opens the more upon me, and I cannot yet get sight of any end of it. The title of the Chapter will be, Of the Conduct of the Understanding, which, if I shall pursue, as far as I imagine it will reach, will, I conclude, make the largest Chapter of my Essay.'51 Locke drafted many pages in working out this new matter for the Essay, but he never completed his account of it and it was not published in any Edition.

It is plain, then, that from March 1694/5 onwards, Locke continued to revise the Essay. As had been done before for the purpose of the Second Edition, the whole book was again carefully examined pg xxixwith a view to making material or formal improvements, and wherever it seemed desirable it was amended. Most of the material amendments were of small extent, consisting of not more than a few words; many, however, were longer and, as already indicated, two additions ran to chapter length. Predominant among the formal amendments were numerous additional italicizations and changes of pointing: Locke carried further, in both respects, his revision for the Second Edition. The over-all tendency of the successive revisions of the pointings in the Essay was towards the greater use of heavier punctuation (e.g. more full stops replacing colons or semicolons, more colons replacing semicolons, more semicolons replacing commas).52 At the same time, until 1699, he was engaged in a prolonged defence of the Essay's doctrines against attacks made on them, and then on his defence of them, by Edward Stillingfleet (1635–99), Bishop of Worcester. Locke's rejoinders to Stillingfleet were published in 1697–9 under the titles A Letter to the Right Rev. Edward Ld. Bishop of Worcester ...; Mr. Locke's Reply to the Right Rev. The Lord Bishop of Worcester's Answer to his Letter; and Mr. Locke's Reply to the Right Rev. The Lord Bishop of Worcester's Answer to his Second Letter. The length of these three publications amounted in total to about two-thirds of that of the Essay. (In 1694–7 Locke was also busy with his Reasonableness of Christianity and its vindications, and with other matters.)

The Third Edition was sold out within three years after its first appearance;53 the book went to the press again in late June or July 1699—Locke stayed in London for the summer and autumn that year—for the Fourth Edition, published around early December 1699.54 This was the last Edition of the Essay in English to appear in Locke's lifetime.

pg xxxThe printers' copy for the Fourth Edition consisted in part of corrected pages of the Second Edition and in part—especially for the longer new additions—of manuscript pages.55 That the Fourth Edition derives principally from the Second (and not the Third) Edition is demonstrated by evidence such as the following: (a) Several formal differences per page in respect of capital letters, commas, apostrophes, or hyphenation usually occur between the Second and Third Editions; the Fourth Edition very largely (though not uniformly) follows the Second rather than the Third Edition in these minor variants. For example: on p. 232 of the Second and Third Editions are eight formal differences, which (with the Second Edition reading being given first) are 'humane'/'human', 'Hypothesis:'/'Hypothesis:', 'sufficient,'/'sufficient', 'same:'/'same:', 'Being,'/ 'Being', 'thing'/'Thing', 'Name'/'name', [twice], 'Ideas,'/'Ideas'; the Fourth Edition shares the reading of the Second Edition in all these cases, except the two involving a colon in italic (where it is the same as the Third). (b) A considerable number of summaries of chapter sections placed in the margin in the Second and later Editions were intended to apply to two or more successive sections; many of these summaries occurred in the margin in the Second Edition at the beginning of two or more of these latter sections. Nearly all such marginal repetitions were eliminated from the Third and later Editions, but a few of them—not the same ones—were retained, by oversight: some marginal summaries repeated in the Second Edition are also repeated at the same place in the Fourth (and, indeed, Fifth) Edition even though they had already been discarded in the Third Edition. For example, the summary for Bk. II, Ch. xxiii, §§ 33, 34, and 35 is printed in the margin at § 33 and at § 35 in both the Second and Fourth Editions, whereas it appears in the margin only at § 33 in the Third Edition; a similar situation is found in Bk. IV, Ch. iv, §§ 2 and 3, as well as elsewhere.

The Fourth Edition was not as accurately printed as the Second, but it was only a little inferior to the Second materially: unlike the Third and Fifth Editions (despite the Errata page of the pg xxxilatter), the Second and Fourth Editions avoid worse than a very-occasional misprint that disturbs the sense. I am sure Locke did some proof-reading of the Fourth Edition. If he was aware of the defects perpetrated in the production of the Third Edition, he would have been especially insistent on the printers avoiding errors this time, and would, then, have thoroughly checked at least portions of their work. I estimate that about 800 copies were printed of the Fourth Edition. The larger additions were again, as in the case of the Second Edition, also 'printed by themselves' for the benefit of owners of a copy of an earlier Edition who wished to bring it up to date with the author's latest thoughts.

The Fourth Edition was published in the form of a folio volume, with the following contents after the title-page: The Epistle Dedicatory (4 unnumbered pages); The Epistle to the Reader (10 unnumbered pages); The Contents [headings of books and chapters] (2 unnumbered pages); The Contents [headings of chapters and summaries of contents of chapter sections] (21 unnumbered pages); Bk. I, Chs. i–iv (pp. 1–39); Bk. II, Chs. i–xxxiii (pp. 41–226); Bk. III, Chs. i–xi (pp. 233–311); Bk. IV, Chs. i–xxi (pp. 313–438); Index (11 unnumbered pages); Errata (unnumbered page; 47 entries). There are seven mispaginations: 91 is numbered 93; 92, 94; 93) 95; 94) 96;56 371, 317; 380, 390; and 381, 391. As in the first three Editions, the pages of the main text generally contained 54 lines, with about 13 words per line; each page had a running (normally, chapter) title, and a catchword at the end.

## (5) The Fifth Edition, 1706

I estimate that about 900 copies were printed of the Fifth Edition.

The Fifth Edition was published in the form of a folio volume, with the following contents after the title-page: The Epistle Dedicatory (4 unnumbered pages); The Epistle to the Reader (7 unnumbered pages); The Contents [headings of books and chapters] (2 unnumbered pages); The Contents [headings of chapters and summaries of contents of chapter sections] (19 unnumbered pages); Bk. I, Chs. i–iv (pp. 1–50); Bk. II, Chs. i–xxxiii (pp. 51–284); Bk. III, Chs. i–xi (pp. 345–440) [sic]; Bk. IV, Chs. i–xxi (pp. 441–604); Index (9 unnumbered pages); Errata (unnumbered page; 83 entries). There are mispaginations at pp. 211, 233, 259–62, [285], 423, 425, 465, 467. Generally, the pages of the main text contained 49 lines, with about 12 words per line; each page had a running (normally, chapter) title, and a catchword at the end.

## (6) The Sixth Edition, 1710

The sixth edition was basically a reprint of the Fifth, with corrections of some misprints of the latter. It marks a new departure only in being the first English edition of the Essay to appear in other than a single volume in folio: it was now published in two volumes in octavo. A. and J. Churchill and Samuel Manship were still the joint publishers of the book.

## (7) The Essay in the Collected Works, 1714

In the first volume of an edition in folio of Locke's works whose title-page runs: 'THE / WORKS / OF / John Locke Esq; [rule] In Three Volumes. [rule] THE CONTENTS of which follow in the next Leaf. [rule] With ALPHABETICAL TABLES, [rule] VOL. I. [rule] [wreath] [rule] LONDON, / Printed for JOHN CHURCHILL at the Black Swan in Pater-Noster-Row, and SAM. MANSHIP at the Ship in / Cornhil. M.DCC.XTV', the Essay was reprinted on pp. iii–xxviii, 1–342, followed by Locke's Letter, Reply, and Second Reply to Stillingfleet, with an Index to the whole volume at the end of it. (Manship's name was not included in the imprint of the second and third volumes of this three-volumed edition.) This edition of the Essay contributed a group of ten words in The Epistle to the pg xxxivReader; they belonged to a paragraph which was new to the Fifth Edition; they had been accidentally omitted from the 1706 and 1710 editions.58

The formal features of this version of the Essay are often different from those of preceding editions, for example in pointing and spelling (e.g. 'Knowledge' is here regularly printed 'Knowledg').

After this edition, the Essay was reprinted with formal modifications on average every three or four years (either on its own or as a part of an edition of Locke's collected works) until at least 1830; and there have been many reprints since that date. Perhaps no other modern work of discursive prose has sold so well and steadily in the course of centuries.

## (8) Coste's French Translation, 1700

Pierre Coste (1668–1747) began translating the Essay into French in July or August 1696. He embarked on this task at the instigation of Le Clerc, who had urged Locke the previous summer to approve this step; Coste's capabilities were already known to Locke, whose Some Thoughts Concerning Education had recently been translated into French by him. For a year or more Coste's progress on the new task went slowly; but the work advanced more satisfactorily when he changed his location: in August 1697 he moved to Oates to become a resident tutor in the Masham household in which Locke lived, and he remained there till after Locke's death. His translation of the Essay was published in about June 1700 (the Privilege is dated 5 May 1700; and see n. 59); it was (piratically) reprinted in 1723 and a revised edition was issued in 1729 which was reprinted half a dozen times in the next half-century.

Coste's translation of the Essay was made with Locke's cooperation, and received his approval announced in a prefatory notice 'Monsieur Locke au Libraire' ('je puis dire au Lecteur que je présume qu'il trouvera dans cet Ouvrage toutes les qualitez qu'on peut desirer dans une bonne Traduction'). In the 'Avertissement du Traducteur' Coste wrote:

My greatest difficulty has been in entering properly [bien entrer] into the author's thought; and despite all my efforts [application], I should often have fallen short [demeuré court] without the help of Mr. Locke, who has been kind enough to revise [revoir] my translation.

pg xxxvAnd in the final paragraph of his 'Avertissement' Coste added:

Although Mr. Locke has courteously given public witness to his approval of my translation, I declare that I do not intend to take advantage of this approval, which at the most signifies that I have broadly caught his meaning [en gros je suis entré dans son sens] but does not at all offer a guarantee against particular mistakes that might have escaped me. Although as I have already said Mr. Locke has listened to the reading of [ait ouï lire] my translation before it was sent to the printer, this does not rule out the possibility of his having let pass many expressions which do not render his thought exactly.

Coste's translation copies the pattern of sentences and paragraphs in the original; the only important rearrangement was the removal of Chapter I of Book I from this Book and the classification of this chapter as an Introduction to the Essay as a whole. His translation is, almost everywhere, admirably clear, readable, and reliable. His procedure as a translator did not, however, aim at a slavish, 'literal' translation, which, indeed, he regarded as impossible in this case and undesirable in principle: the original provides the translator with a message which it is his responsibility, like an ambassador, using appropriate discretion, to convey in another language ('il me semble que le Traducteur et le Plenipotentiaire ne sauraient bien profiter de tous leurs avantages, si leurs Pouvoirs sont trop limitez'). Very frequently, Coste's version gives a more explicit articulation of what he gauged to be Locke's meaning. Cf. 433(14), n.

From what original form of the Essay did Coste make his translation? In the title-page of his first edition and in a note in the translated Epistle to the Reader in his revised version of 1729, Coste said that his version of 1700 was made from the Fourth Edition. The Fourth Edition had not been printed, nor even its preparation completed, during at least almost the whole time Coste was translating the Essay; but his version did include most (though not all) of the changes from earlier Editions that Locke made for incorporation in the Fourth Edition: Coste's version represents an intermediate phase, prior to that of the Fourth Edition. The text of the Essay on which Coste chiefly relied was that of the Third Edition. This is shown by his adopting several misprinted readings peculiar to the Third Edition. For example, he writes 'ils sont' for 'they are', which only the Third Edition misprinted for 'they were' (Bk. II, Ch. xiii, pg xxxvi§ 9); 'pour le mouvement' for 'for the Motion', which only that Edition misprinted for 'from the Motion' (Bk. II, Ch. xvii, § 4); 'ceux qui l'écoutent' for 'Hearers', which only that Edition misprinted for 'Hearer' (Bk. III, Ch. ii, § 2); 'indifféremment' for 'indifferently', which only that Edition misprinted for 'differently' (Bk. III, Ch. ix, § 17); 'chaque' for 'every', which only that Edition misprinted for 'very' (Bk. IV, Ch. iii, § 19); and 'Mais' for 'But', which only that Edition misprinted for 'Put' (Bk. IV, Ch. vi, § 11). Such errors passed unnoticed by Locke because, as both he and Coste remarked in their prefatory notices to the published translation, Coste had read the translation to him: Locke did not make a word by word comparative examination of translation and original. Coste also copies some formal features peculiar to the Third Edition (see e.g. Bk. IV, Ch. x, ad fin.).

Apart from some typographical details, the title on the title-page of Coste's version of 1700 runs: 'ESSAI PHILOSOPHIQUE / CONCERNANT / L'ENTENDEMENT / HUMAIN, / OU L'ON MONTRE QUELLE EST L'ETENDUE DE NOS / CONNOISSANCES CERTAINES, ET LA MANIE'RE / DONT NOUS Y PARVENONS. / Traduit de l'Anglois de Mr. LOCKE, / Par PIERRE COSTE, / Sur la Quatrième Edition, revûë, corrigée, & augmentée par l'Auteur.' The title on the title-page of Coste's revised version of 1729 is the same down to 'PARVENONS.', and then continues: 'PAR M. LOCKE. / TRADUIT DE L'ANGLOIS / PAR M. COSTE. / Seconde Edition, revûë, corrigée, & augmentée de quelques Additions / importantes de l'Auteur qui n'ont paru qu'après sa mort, & de quelques / Remarques du Traducteur.' In both editions, only the epigraph from Cicero, which had appeared on the title-page of the first three Editions in English, was displayed; the other epigraph, from Ecclesiastes,—first added to the title-page in the Fourth Edition—was not included. (Both epigraphs are on the title-page of Burridge's Latin translation of the Essay.) Amsterdam was the place of publication in 1700 and 1729.59

## pg xxxvii(9) Burridge's Latin Translation, 1701

Le Clerc had promised, as early as 1688 or 1689, to translate the Essay into Latin. After the Essay had been published, an early initiative was taken by a young man (Verryn) who wrote to Locke offering to turn the book into Latin; Locke eventually declined Verryn's translation because it was insufficiently competent. Molyneux tried to obtain the services of a man called Mullart to undertake a Latin translation of the Essay; it was plain after a short time, however, that Mullart's other commitments and interests would prevent him from giving effect to his preliminary agreement with Molyneux to produce such a translation. As soon as this plan had fallen through, in 1695, Molyneux invited another acquaintance of his in Ireland, Ezekiel Burridge (1661–1707), to make a Latin translation of the Essay; Burridge accepted. At first, it looked as if his translation was making little headway, but by 1698 the prospect of its being soon completed began to seem promising. Burridge's (anonymous) translation was published by A. and J. Churchill in 1701; it was reprinted on the Continent several times in the eighteenth century. Although not nearly as much as Coste's French translation, this Latin version—most of whose copies probably went abroad—contributed to the wider diffusion of Locke's epistemology in Europe.

According to its title-page, the text of this translation was based on the Fourth Edition ('De Intellectu Humano. / ... / Editio Quarta Aucta & Emendata, & nunc primum / Latine reddita.').60

# C. THE METHODS OF THIS EDITION

## (1) General

As already indicated, I have planned this edition of Locke's Essay and related materials to be in two parts. The present volume incorporates the first and principal part which handles material in whose publication Locke himself had taken an active interest; it is concerned exclusively with the printed versions of the Essay. The further, supplementary publication (Drafts)—apart from containing a general and textual Introduction—will chiefly deal with pg xxxviiimanuscript materials; this supplement is intended to offer (a) transcriptions of all the available manuscripts (other than correspondence)61 that are directly related to the Essay from its embryonic phase starting in 1670 or 1671 till Locke's death in 1704,62 and (J) new editions of certain Lockian writings, also directly related to the Essay, which were originally published with Locke's authority but which have not previously been collated with manuscript drafts of them. The contents of Drafts will, broadly speaking, be in chronological order. Items under (a) include Drafts α,63 A, B, and C of the Essay; agreements with bookseller-publishers, journal entries, etc. relating to the printing or publication history of the Essay; Locke's comments on criticisms of his book; and his intended additions to or other modifications of it. Items under (b) include the Epitome64 and Of The Conduct of the Understanding.65

If there had survived, in whole or in part, the manuscript that was used as printers' copy for the First Edition of the Essay,66 or a close forebear of it, it might have been practicable and desirable to record within a single publication its special readings in comparison with the collated printed versions. In the absence of such a manuscript, it is best, for at least practical reasons, to separate completely the editing and annotation of the published Essay from the treatment of manuscript sources and of subsidiary publications.

The remainder of this Introduction is solely about the methods applied to the materials covered by the present work. From now pg xxxixon, for the sake of clearness, 'Text' will be used to mean the text of the Essay that is established in this volume.

## (2) Choice and Treatment of Copy-Text

Although the First Edition could, with some historical advantage, be chosen to serve as copy-text, its use for this purpose would have—on balance—the greater drawback of excluding a text that incorporates Locke's revisions. Nor could the latter text be exactly reconstructed by the reader from information in a critical apparatus taken in conjunction with a reproduction of the First Edition, unless the critical apparatus was so complete for every detail of spelling, hyphenation, apostrophization, punctuation, italicization, capitalization, etc., as well as of wording, as would surely make the task of reconstruction excessively hard and troublesome; it would also require the critical apparatus to be, for most readers, too heavily and disproportionately occupied by fine points, such as every change in spelling or in the occurrence of a capital letter, an apostrophe, or a hyphen. Hence, the Edition to be chosen as the copy-text should at any rate be one that is late in the sequence of the Editions that Locke authorized. Admittedly, for the very reasons I have just been mentioning, this inversely entails in practice the impossibility of reconstructing every feature of the text of the First Edition (but see, further, p. lii below); a reader interested in features of the First Edition that are not covered by the Text or annotations of the present edition will have to examine that Edition in the original or in a photo-reproduction.67

The posthumous Fifth Edition, whose publication had been authorized by Locke shortly before his death,68 would then prima facie seem to have the strongest claim to be chosen as copy-text. And this claim would have decisive force but for the existence of two counteracting considerations. Firstly, Locke did not live to begin to see the Fifth Edition through the press, whereas in the case of preceding (especially the First, Second, and Fourth) Editions he had tried to ensure that his book as published would represent his latest wishes even in its formal aspects, e.g. its pg xlpunctuation and italics.69 Secondly, the formal aspects of the Fifth Edition are only very approximately the same as those of the Fourth, as the register of formal variants included below makes evident. The incidence of initial capital letters for nouns in nearly all of the first half, and in the Index, of the Fifth Edition is a great deal more common than it is in the preceding Editions; the tendency in those parts of the Fifth Edition is for all nouns except ones occurring as minor concomitants to be initially capitalized,70 and indeed a similar capitalization of adjectives and verbs is also more frequently found there than it is in the first four Editions. In both these respects the Fifth Edition departs far less often from its predecessors in Books III–IV. This applies likewise to spelling, apostrophization, and hyphenation; and, more important, to punctuation. In regard to the last it should be emphasized that pointings in the first half of the Fifth Edition often differ from those in the Fourth, that many are certainly non-Lockian or erroneous (e.g. at 336(11), 549 (24), 554(6), on which see the register ad loc.), and that claims that the special pointings of the Fifth Edition are a late systematic revision by Locke himself should be received with caution until a careful comparative study of punctuation in the Essay and his other works has been undertaken.71 My view is that only a small number of the posthumous pointings are acceptable at present as genuinely Lockian in origin or as corrections in accord with his preferred style in print. Further, the Fifth Edition has many more dubious readings, that are probably if not certainly misprints, than the pg xliFourth Edition has. Because of these considerations I have thought it safest to take the Fourth Edition as providing the copy-text here.72

However, the present Text of the Essay is not based only on the copy-text. A number of sentences or even longer passages were first printed in the Fifth Edition, which also introduced many changes of a word or a few words; it also accidentally made some omissions, e.g. of the passages at246(10–12),415(12–13),419(5–7), and515(10–11). (Early, and recent, editions derived from the Fifth Edition wrongly omit these passages.) The readings of the Fifth Edition have been incorporated within the Text wherever I have judged them to be probably authentic—and not misreadings or misprintings. Of course, no code exists to guarantee correct or objective textual decisions; an editor has to exercise his discretion in each individual case guided by his general knowledge, experience of the texts, and a consideration of the context.73 This does not imply that an editor could not state the factors which he has weighed up in making a textual decision; it is because some of the factors are incapable of being measurably or objectively weighted in relation to one another. Nor does it imply that his decisions cannot rationally be discussed and criticized. Sometimes confidence, let alone certainty, about what is the best reading for the Text has been far from attainable; a preference may have emerged although admittedly being very conjectural, or alternative readings may persist in seeming to he more or less equally defensible.74

Thus to form, for a critical edition, an eclectic text that is principally the same as that in an edition published in the author's pg xliilife-time but modified to incorporate material first published post-humously, is by no means an unprecedented or irregular procedure; it has been adopted for, e.g., many Shakespearian plays. One notable parallel in this way to the Essay is Montaigne's Essais (for which, however, the author's extensive amendments to the last edition (1588) published in his lifetime that first appeared only posthumously (1595) have survived in their original manuscript form in 'l'exemplaire de Bordeaux').

I have not included in the Text or the critical apparatus the Locke-Stillingfleet material introduced in footnotes in the Fifth Edition. It is uncertain whether Locke himself wished thus to add to the Essay, nor is there any evidence—somewhat on the contrary, since the prefaces to these footnotes speak of him in the third person—that he played any part in their selection and presentation. This material in the footnotes is not part-and-parcel of the Essay, but a series of rather one-sided excerpts from other relevant publications. Further, although only excerpts were given in the Fifth Edition, they occupied a total of many pages, and to include them here would both considerably enlarge this volume and yet be of little service to the serious scholar, who would perforce still have to examine the complete, original publications in the Locke-Stillingfleet controversy. I have, however, in the critical apparatus noted these additions to the Fifth Edition, quoted the prefatory remarks that accompanied them, and cited the references given in the margin of that Edition;75 and cf. Appendix, Note B.

The material from the body of the Fifth Edition in the Text has not always been reproduced in its original form. I have amended the formal features of the material that is inserted from that Edition to make it more uniform with its immediate surroundings in the Text. I have annotated these amendments and cited the original forms in all cases in the register of formal variants.

The character of the copy-text has been retained as far as possible, with some minor exceptions. Apart from the incorporation of Fifth Edition material, the Text is not an exact reproduction of the copy-text in layout or content in the following respects. (i) Some readings of the Fourth Edition seem to require correction, in the light of other early printed versions of the Essay or of other considerations, pg xliiieither in such small things as reinstating a comma, introducing a paragraph break, or replacing a lower-case by an upper-case letter; or in materially altering a word. I have kept such alterations within strict bounds, doubtless at the cost of complete consistency. (ii) The layout is slightly different, especially in that the summaries of the chapter sections that were printed in the margin from the Second Edition onwards, and Locke's marginal references to sources (to a few of which I have added a reference symbol lacking in the original), have here been placed immediately under the body of the Text, and the relevant section numbers have been prefixed to the summaries. (iii) Duplications of the original marginal summaries have been silently eliminated.76 (iv) Trivial misprints, e.g. an upturned letter, an accidental misspelling such as 'Schalostique' for 'Scholastique', or an italicized letter in a word in roman type, have been corrected in the Text and annotated in the register of formal variants. (v) Some contractions, e.g. 'CHAP.', 'SECT.', '&' (and '&'), and 'Aq.' have been expanded. (vi) Some minor typographical changes have been made in a few places; e.g. '*' has been replaced by '†', and italic punctuation after italic words in passages printed otherwise in roman type has been replaced by roman punctuation. (vii) The full stop at the end of the running book and chapter numerations and running titles, of some other headings, and of § has been discarded. (viii) The long 's' has been replaced and several ligatures have been disregarded. (ix) The amount of matter on a page of the Text bears no relation to that on a page of the copy-text; nor are the original paginations or catchwords mentioned in the Text or critical apparatus. (x) In the Index, the paginations of the copy-text have been replaced by ones adapted to the present edition; also, where it is appropriate, the original 'ibid.' or 'ib.' has been replaced by the new pagination. The section symbol 'S.' has been replaced by '§' and 'ib.' by 'ibid.' The numerous occurrences of 'p.' have been omitted. Some incidental italicizations have been romanized. No reference has been made to Coste's Index. Formal variants in the Index have not been annotated; and I have disregarded the fact that it was printed in italic in the Second and Third Editions. (xi) The Text contains some editorial signs, especially asterisks referring to sources of quotations, and italicized expressions enclosed within square brackets, e.g. '[bis]'.

## pg xliv (3) Material Variants

The critical apparatus subjacent to the Text chiefly aims to record all items belonging to one or other of two (overlapping) categories of readings. One category is of readings in the Text that do not coincide materially, or formally,77 with the copy-text (the Fourth Edition). The great majority of the entries in the critical apparatus record readings in the other category: where the wording is not identical in all the early printed English versions, from the First Edition to the Fifth Edition inclusive; or where there is a change of pointing that is semantically considerable or that seems to have been made in order to prevent an ambiguity or misinterpretation.

The main editorial symbols and abbreviations, with their meanings, used in the annotations in the critical apparatus are the following; examples of usage are given on pp. xlvii–xlviii.

 1 the First Edition78 2 the Second Edition 3 the Third Edition 4 the Fourth Edition 5 the Fifth Edition 6 the sixth edition W The Works of John Locke (1st edition), Volume I Coste Coste's translation (1st edition) Coste2 Coste's translation (2nd edition) pg xlv er Errata (page in 1, 2, 4, or 5; indicated by e.g. '2er') 1T.er manuscript additions to the Errata page in Tyrrell's copy of the First Edition (British Museum, C. 122, f. 14) add. added edit. editorial om. omitted; this is used in stating an absence from a later version compared with the content of an earlier version ('not in' is used in stating an absence from an earlier version compared with the content of a later version) ] end of quoted expression from Text / end-of-line marker | separation sign between successive annotations of a quoted expression from Text

Some further signs are employed in the register of formal variants:

 ital. in italic type rom. in roman type ~ repetition of guide-word(s) ∏ initial capital letter ∩ initial lower-case letter

Editorial wordings (including abbreviations, except 'rom.') in annotations are, for the sake of distinctness, printed in italic type. This has sometimes led me to use roman instead of italic type in citing titles of books and other matter.

Unless the contrary is indicated by an annotation in the critical apparatus, (i) the wording of every portion of the Text other than the marginal sectional summaries is to be assumed to be identical (apart from trivial misprints, and, possibly, formal features) in the first five Editions; and (ii) the reading in the first edition of Coste's translation is normally to be assumed to be semantically equivalent to (even if more explicit79 than) the reading in the Fourth Edition. If the Coste version is markedly closer in sense to one or more other Editions than it is to the copy-text, '(likewise Coste)' is appended to the annotation of the former; if the Coste version does not take account of an addition made in the Fourth Edition, this is usually noted by '(Not in Coste)'. It needs to be borne in mind, however, pg xlvithat the Coste version is sometimes, in effect, neutral—or amphoteric—owing to differences between English and French or between Locke's style in the Essay and Coste's style in his translation of the book.'80

The marginal sectional summaries of the copy-text that are printed here immediately below the rest of the Text are annotated along with the latter; but they are not repeatedly noted as having been added subsequently to the First Edition. Each such sectional summary is to be assumed as having appeared, with the given wording, from the Second Edition onwards.

If a quoted reading in a note in the critical apparatus occurs in two or more successive Editions, an en rule in bold type (';-') is used between the numerals in bold type denoting the first and last of these Editions; where there is a break in the succession before a renewal of the reading, a comma is used between the numerals denoting the Editions. But it is important to appreciate two points of interpretation in all such cases. (i) Where reference is made to the Fourth Edition as a source of a quotation, the quotation is an exact one both materially and formally in respect of that Edition; it is intended to be exact materially, but not necessarily formally, in respect of the other Editions referred to in common with the Fourth. (ii) Where reference is made to the First but not to the Fourth Edition as a source of a quotation, it is an exact one both materially and formally in respect of the First Edition; it is intended to be exact materially, but not necessarily formally, in respect of the other Editions referred to in common with the First. (In fact, broadly speaking, any quoted passage common to the first four Editions is very much the same in them in formal respects.) Minor formal differences between the Editions are noted in the register of formal variants.

Generally in the critical apparatus, quotations from the Text do not include any item of punctuation that immediately follows the last quoted word unless such a final pointing is different or absent from the contrasted reading in another Edition; thus, if a wording quoted from the Text has its original source in the Fourth Edition where it is followed by (e.g.) a colon and the wording in the first three Editions is unlike that but is also followed by a colon, then, pg xlviias a rule, this shared item of punctuation is not printed in the annotation and the reader is left to supply the terminal pointing (of a colon in the given example) to the wording quoted from the first three Editions on the basis of the terminal pointing printed in the Text. (In the event of a clash between the preceding paragraph and this one, the former takes precedence.)

Some annotations in the critical apparatus are further qualified there by means of an additional note placed within brackets; this most often arises where two or more Editions that are grouped together as sharing a reading differ in some particular respect.

An amended reading given in the Errata page of 1, 2, 4, or 5 is sometimes treated as the reading in the Edition and the original reading is not separately noted; an Errata correction that merely restores a reading given in all preceding Editions is not registered. As a general rule, new differences introduced in an Errata page are recorded, and are distinguished by using the suffix 'er' with the numeral for the Edition; and the original reading in the Edition is also stated.

Misprints not listed in Errata pages are not noted if they are trivial formal ones, except in the case of the Fourth Edition. The numerous 'verbal exchange' misprints,81 i.e. erroneous substitutions of words by other words (e.g., 'sense' for 'Scene', see 154(4), n.; 'Word' for 'World', see 163(30), n.; 'Letter' for 'Letters', see 175(23), n.; 'for' for 'from', see 211(20), n.), are usually noted in the critical apparatus; but some verbal misprints, especially those in the Third Edition in which they occur frequently, are recorded in the register of formal variants.

Running titles, running numerations, and catchwords have not been annotated.

Annotated matter is not marked by any reference symbol in the Text; each editorial note is simply prefixed by the relevant line number(s). But superscript numerals are used in annotating the Contents and the Index, and very occasionally elsewhere. A note is inserted in the critical apparatus to signal the end of an annotated passage which runs in the Text from one page to another.

Some examples to illustrate the method of annotation in the critical apparatus are:

P. 129: '(32) this] 25 | the other 1'

pg xlviiiThis means that, in l. 32 on p. 129, the wording in the Second, Third, Fourth, and Fifth Editions is 'this' while 'the other' is the wording in the First Edition. Further, because 'this' in the Text is followed by a colon, it is legitimate to infer that the Fourth Edition reads 'this:' and that the First Edition reads 'the other:'.

P. 225: '(7) [1st] our] 1er5 | these 1'

This means that, for the first occurrence of 'our' in l. 7 on p. 225, the Errata page in the First Edition gives 'our'; that the Second, Third, Fourth, and Fifth Editions read 'our'; and that 'our' is the replacement of 'these', which is the reading in the text of the First Edition.

P. 358: '(21) which ... in] 45 | of 13'

This means that the wording 'which are to be found in' in l. 21 on p. 358 is in the Fourth and Fifth Editions; and that this wording is the replacement of 'of' in the First, Second, and Third Editions.

P. 353: '(24)–354(2) For ... Vice.] add. 25'

This means that the passage in the Text running from 'For' in l. 24 on p. 353 to 'Vice.' in l. 2 on p. 354 is in the Fourth and (apart possibly from its formal features) in the Second, Third, and Fifth Editions, but is not in the First Edition.

P. 404: '(17–18) are; ... consist;] 4er5 | are, ... consist, 14'

This means that the reading 'are; wherein they consist;' in ll. 17–18 on p. 404 is based on the Fourth Edition's Errata page (which has for this place 'are;' and 'consist;' as two entries) and the text in the Fifth Edition, whereas the reading in the text in the first four Editions is (apart possibly from formal features other than the noted pointings) 'are, wherein they consist,'.

P. 136: ';(7) Brains or] 25 (Brains, 5) | Brains, 1'

This means that 'Brains or' in l. 7 on p. 136 is the reading in the Second, Third, and Fourth Editions; that 'Brains, or' is the reading in the Fifth Edition; and that 'Brains,' is the reading in the First Edition.

## (4) Formal Variants

The idea still persists that English printers in the seventeenth century took little or no notice of the formal features of an author's pg xlixwork as displayed in his manuscript, and that the pointings and other formal features of a book were predominantly if not wholly determined by the practice of the individual printing-house or compositor. A natural correlate of this idea is that changes in the formal features of a book in later editions were always the result, not of an author's wishes, but of printers' decisions.82

This sweeping assertion is largely correct for some books of the period, especially where authors did not declare in advance their requirement that the formal features of their manuscript be followed in print. But it is false overall as is shown by the attentive rendering of surviving copy in the first editions of a number of works, e.g. Paradise Lost, Book I, and Newton's Principia,83 and by the fact that some authors went to much trouble to mark in the margin of a copy of an edition their desired changes in formal, as well as material, respects for the purposes of a new edition—which they would hardly have done if it was expected that the printer of a new edition would quite disregard an author's revisions of this kind.

It seems to me that if one considers a broad spectrum of literature of this period, it is clear that a more limited role was played by the printer in modifying copy: by far the greater number of a manuscript's pointings, capitalizations, apostrophes, and indications of italics, were adopted; on the other hand, unless he was firmly directed otherwise, the printer often took the liberty of inserting punctuation where pointings appeared too sparse, of increasing the capitalizations, apostrophes, and italicizations, and of printing more modern and less variegated spellings than those used by the author. It has also to be recognized that probably pg lnearly all authors expected and accepted that the printers would exercise this liberty.84 As for authorized editions subsequent to the first, it is reasonable to presume, on business grounds, that if an author was fussy in wanting certain changes in formal features of his book, his preferences would be put into effect.

Let me turn from these generalities to Locke's Essay. It might be maintained that no annotations of the formal aspect of the published Essay need or should be given because Locke was in no way concerned with, or responsible for, that aspect in any of the printed versions. However, this would be a conclusion resting on the untenable assumption that the formal features of every book produced in England in the seventeenth century were settled by the printing-house and not at all by the author. Alternatively, it might be suggested that notice should be paid only to formal variants which are significant changes of punctuation. But this would lead to a too limited and unsystematic treatment of the formal aspect, since significant changes of punctuation constitute a narrower class than changes of a formal kind that make a syntactical or semantic difference, or whose appearance was probably at Locke's instructions; and nothing less—and, for the historical record, more—than the latter ranges should be noted in a critical-text edition.

The contents (apart from printers' errors), in wording and form, of the successive editions of the Essay that were published while Locke was alive can and should be regarded as having had his approval at the relevant time. What Locke thus approved is not to be identified in every detail with what he (or an amanuensis) had written for publication, for the compositors introduced some minor alterations in formal respects; but in general these would not have been retained without Locke's consent.

Numerous changes in formal features of the Essay were made in the Second Edition; among these changes were alterations of punctuation. Tyrrell's copy of the First Edition of the Essay has fifty-two handwritten additions to the Errata page that seem to represent a very early revision of the First Edition; that they are authentically Lockian is confirmed by the fact that they almost completely correspond to revisions to be found in the text of the Second Edition; a considerable proportion of them solely affect pg lipunctuation.85 For the reason just given and in view of the care bestowed by Locke and Molyneux (and possibly other friends of Locke's) on detecting even small errors in the First Edition, it is out of the question that any of the various alterations inscribed in the Errata in Tyrrell's copy derives from the printers' initiative. Likewise, Locke himself must be the source of such extensive formal changes occurring in the Second Edition as the many new italicizations, and the majority of the many modifications of pointing (but some of the omissions or additions of commas doubtless result from compositorial practice). The Third Edition was almost identical in stronger pointings and in italicizations—and closely similar in other formal respects—with the Second; this confirms that the printers of the first four Editions of the Essay did not on their own account tend to make formal changes other than occasional very minor ones.

A large number of further formal changes were introduced in the Fourth Edition; again, the responsibility for the many new italicizations and the majority of the many modifications of pointing was almost certainly Locke's. It would have required a compositor of exceptional acumen and conscientiousness, willing to expend a good deal of unrewarding labour, to have initiated them; and anyway such a man would not have done this, at the risk of imperilling future business, without appropriate approval: the publisher knew very well that Locke was a fastidious author.

Further evidence of Locke's own care about the correction and improvement of the formal aspects of his writings is provided by the late revisions of his Two Treatises of Government. 'The better quality reprint [of the Two Treatises of Government] was issued as Locke had demanded, the 3rd edition, 1698 ... But even this did not satisfy Locke, who seems to have had a standard of perfection above the resources of the printers of his time. This 3rd printing of 1698 had its faults, but it is difficult not to feel that the exasperation which he showed in his will over all the printings of this work had an independent source in an inner anxiety about what he had written. As it became obvious to him that no version correct enough to satisfy such meticulousness would ever appear in his lifetime, he made plans to ensure that it should do so after his death.

pg liiHe corrected a copy of the printed version in minute detail, scrutinizing the word-order, the italics, the punctuation, even the spelling, as well as the general sense. It seems that he intended to carry out this process in duplicate, which is what we might expect in him.'86

There are critical-text editions of English literary works of the sixteenth or seventeenth century which recognize a standard of annotation that entails recording all variants of a material or formal nature, down to every jot or tittle.87 I accept this in principle as the proper standard for an ideal critical-text edition of the Essay. I have, nevertheless, lowered my sights in the present edition in relation to the formal variants. The inclusion in the register of formal variants of every single detail might, by its very extent and completeness, discourage attention from, rather than attract it to, the formal aspect of the Essay; nor have I wanted to let the register become so full as to swell this volume beyond manageable proportions. The register is, however, intended to provide as comprehensive an account of the formal variants as should satisfy all students of the Essay and its history unless their interests in the formal variants could not well be met by anything less than their making a minutely close and thorough examination of the early printed versions individually.

The register records variants not already noted in the critical apparatus.88 Subject to two qualifications to be mentioned, the annotations cover a majority sample of the formal variants arising from the early Editions. The degree of the register's inclusiveness varies somewhat; but for Book II onwards it is commonly at least seventy per cent per page. Formal variants, in this context, consist of differences between the Editions in the use of punctuation, parentheses, roman or italic type, spelling, apostrophe, upper or lower case, word-division, and line-division.

pg liii(i) Line-divisions are not systematically treated in my notes; and (ii) my coverage in regard to upper or lower case is comprehensive only for the 1690–1700 Editions and for variants within—not (e.g. after a colon) at the beginning of—a clause or sentence.

The method of annotation in the register of formal variants is similar to that in the critical apparatus. Some additional abbreviative devices are employed in it. Firstly, 'ital.' means that the cited word or whole group of words is in italic type in the enumerated Edition or Editions; for example, '293(4) Ideas] 25; ital. 1' signifies that 'Ideas' in l. 4 on p. 293 is the reading in the Second, Third, Fourth, and Fifth Editions, while 'Ideas' is the reading in the First Edition. Secondly, 'rom.' means that the cited word or whole group of words is in roman type in the enumerated Edition or Editions; for example, '168(3–4) Feet ... Fathoms] 25; rom. 1' signifies that the reading 'Feet, Yards, or Fathoms' in ll. 3–4 on p. 168 is in the Second, Third, Fourth, and Fifth Editions, while 'Feet, Yards, or Fathoms' is the reading in the First Edition. Thirdly, three other signs are used, especially in treating Bk. II, Ch. xxii onwards where the intensity of formal variants sufficiently warrants their adoption: '~' represents the corresponding guide-word(s) apart from any adjacent punctuation; '220F' standing on its own, or prefixed to '~', signifies that—prescinding from any words covered by ellipses in the guide-expression—each affected guide-word has an initial capital letter, and '∩', marking the occurrence of lower-case letters, is used in ways strictly parallel to those for '∏'. Thus, the latter part of '321(13) Older, and Younger] 14| ∩~ and ∩~ 5' signifies that the relevant reading in the Fifth Edition is 'older and younger'; and the latter part of '333(3) definitions ... observation] edit.; ∏ 5' signifies that the relevant reading in the Fifth Edition is 'Definitions ... Observation'.

## (5) Sources of Quotations

Locke usually states the sources and the locations (i.e. the chapter or page, etc.) of passages he quotes or refers to that are in seventeenth-century works. But he does not indicate any source in a considerable proportion of the passages he cites that are drawn from Classical authors, the Bible, or other older writings; I have, in such cases, added a brief footnote (below the critical apparatus) giving the relevant information.

pg livIn the Essay, from the First Edition onwards, Locke used several different methods of reference to sources. One of these methods calls for a word of elucidation because of its unfamiliarity to modern readers. It consisted of naming an author, or a book's title, together with a fraction, e.g. '$201 322$'; this signified that reference was made to page 201 of the book alluded to or named, which had a total of 322 numbered pages in the edition Locke employed: the numerator of the fraction gave the page number and the denominator gave the last number in the pagination of the book.89 In giving references and cross-references, he made some use of such abbreviated forms as 'l.' and 'c.' derived from the Latin words 'liber' and 'caput' for 'book' and 'chapter', respectively.

For reasons of space as well as for other reasons the only commentary here provided on the Text is a linguistic one, in the form of an extensive Glossary which it is hoped will prove useful to readers unfamiliar with the relevant senses of such expressions in the Essay as are now obsolete or possibly unclear, including Latin and other foreign words. For a fuller account of the Glossary, see its headnote.90

## Notes

1 Essay,704(20–1), i.e. p. 704, lines 20–1, of this edition of the Essay. On Locke's view of the nature and role of reason, sec entries under 'Reason' in the Index to the Essay; see also Carlo Augusto Viano, John Locke: dal razionalismo all'illuminismo (Turin, 1960), especially pp. 327–87, 545–609. Viano's book gives the most broadly erudite and instructive account, and is the most balanced and best organized in its coverage, among existing books on Locke's thought as a whole.

2 I use 'Edition' (with a capital) only in referring to the first five English editions of the Essay (1690–1706), 'version' as a more general term to cover these together with the later English editions of the Essay until the tenth edition of 1730, and the French and Latin translations of the Essay published in 1700–29, and 'early versions' for versions published not later than 1706. The meanings of the standard bibliographical and typographical terms used in this Introduction, and some text-critical topics, are discussed in Note A of the Appendix.

3 Fraser's text is inaccurate owing to his unwarranted practice of frequently adding, omitting, or interchanging words, altering singular nouns into plurals or vice versa, altering tenses, etc. His treatment of Bk. II, Chs. xii–xiii is representative; with his readings that diverge from the original first (and his pagination): I, p. 213, 1. 6: 'one' (with Fifth Edition) for 'any one'; (l. 11 's;others' for 'other'); p. 214, l. 3: 'much' for 'muchwhat'; l. 14: 'together' added by Fraser; p. 215, l. 15: 'as' added by Fraser; l. 21: 'later' for 'latter'; p. 216, l. 3: 'to' for 'in'; l. 28: 'we trace' for 'we will trace'; p. 217, l. 6: 'operations' for 'operation'; (l. 16: 'others' for 'other'); p. 220, l. 21: 'The' for 'This'; p. 221, l. 24: 'join another' for 'join it to another'; l. 27: Fraser's first comma replaces 'or'; p. 222, l. 25: 'these' added by Fraser; p. 223, l. 21: Fraser omits 'properly' after 'said'; p. 224, l. 1: 'part' for 'parts'; l. 15: Fraser omits 'only' after 'know'; l. 17: Fraser omits 'our' before 'use'; p. 228, l. 10: 'thereby' for 'hereby'; p. 229, l. 18: 'spirit' for 'spirits'; l. 21: 'body' for 'a body'; p. 230, l. 19: marginal summary is Fraser's; l. 22: 'is' for 'was'; p. 231, l. 4: 'substantio' for 'substantia'; l. 25: 'and' added by Fraser; p. 232, l. 14: Fraser omits 'as' after 'so'; (p. 233, l. 8: 'seems' for 'seem'); p. 233, l. 20: first 'the' added by Fraser. (I have put in brackets the cases where Fraser's readings are legitimate modernizing changes.) Fraser has throughout altered the original uses of roman and italic.

An illustration of Fraser's confusions about the versions of the Essay is op. cit, I, p. 24. His text for the concluding paragraph of The Epistle to the Reader begins 'In the Sixth Edition there is very little added or altered' and he appends a note on this paragraph which states that the sixth edition was issued in 1706—whereas it was the Fifth Edition that appeared then, the sixth edition being issued in 1710; that the sixth edition 'contains a few slight additions and alterations'—he means (cf. ibid., p. xiv) to the Fourth Edition; and that most of these additions and alterations had 'appeared in Coste's French Version of the Essay'—whereas (a) the 1706 Edition has more than a few departures from the Fourth Edition; (b) not all such departures are slight; (c) it is false that most of the departures had appeared in Coste's French translation of 1700; (d) indeed, although occasionally this translation does conform with the Fifth Edition in contrast with the Fourth, there are also readings in the translation which match the Second or the Third Edition and do not conform with a later one.

Some of Fraser's references to Coste's version would be more consistent and comprehensible if he were using Coste's revised translation of 1729, rather than Coste's first one; for example, he says (op. cit., I, p. 97, n. 7), concerning the sentences in the text that mention La Loubère and Navarette, that these sentences were 'added in Coste's French Version'; in fact, they originally appeared in the Fifth Edition of 1706, were not included in Coste's translation of 1700, but were included in the revised translation of 1729.

4 P. H. Nidditch, 'Corrigenda to Yolton's edition of Locke's Essay', The Locke Newsletter, ed. Roland Hall, 2 (1971), 21–9 and ibid., 3 (1972), 34–8; and A. D. Woozley, ibid., 2 (1971), 29–30 and 3 (1972), 39. See also p. xii below.

5 The modern editions of Draft A and Draft B by R. I. Aaron and Jocelyn Gibb (Oxford, 1936) and by Benjamin Rand (Cambridge, Mass., 1931), respectively, are incomplete through the omission of all such changes, which ought to be included in scholarly transcriptions and are one of the most interesting features of the manuscript material.

6 The dates given here of versions of the Essay are those appearing on their title-page; the actual year of publication may have been different—a point considered later in this Introduction.

The Julian Calendar remained the standard in the British Isles at this period; unless noted as being in New Style, pre-1752 dates in the Introduction accord to the Julian Calendar, which in Locke's time until 28 Feb. 1700 was ten days, and then became eleven days, behind dates in the Gregorian Calendar; in Old Style, the year commenced on 25 Mar. Calendrical references occur in the Essay at e.g. 5(19), 189(11), 192(7), 194(21), 201(3), 636(30); the first of these is certainly in Old Style. The abbreviations 'O.S.' and 'N.S.' are used to mean 'Old Style' and 'New Style' respectively. In referring to the period from 1 Jan. to 24 Mar. in O.S., a convention is to mention both the O.S. and the N.S. year, separating them by '/' as in '1686/7'.

7 For comparisons between Locke's epistemology before and as from 1671, see W. von Leyden's edition of Locke's Essays on the Law of Nature (Oxford, 1954), pp. 60–82; on the development of his epistemology in the years of his preparation of the Essay, see Richard I. Aaron, John Locke, 3rd edn. (Oxford, 1971), pp. 15–73 and Viano, op. cit., pp. 421–609.

8 Draft A (in the ownership of A. A. Houghton, Jr.), page 56 (as originally numbered in the notebook Locke used for this draft).

9 Draft B (Bodleian Library, MS. Locke c. 28, f. 26), unnumbered page, page 1, and page 3, respectively. 'Opinion' was written as a substitution for 'Beleif', which was cancelled. Cf.43(14–16) below.

10 Draft C (Pierpont Morgan Library, New York City), leaf 2, recto. (The leaves are unnumbered in the original; for numbering purposes I disregard the initial blank leaves, and count the page with an inscription to Edward Clarke, James Tyrrell 'or' David Thomas as being leaf I, recto.) This manuscript covers only Bks. I and II. In the quoted title, the date 1685 seems to have been added later.

11 e.g. letters to Pembroke, 8 Dec. 1684, and Furly, 19 Jan. 1688 (n.s.); Tyrrell, in his letter to Locke, (c. 20) Dec. 1689, after publication of the Essay, refers to the book as De Intellectu. Leaf 3, recto, of Draft C, which contains at the top the heading 'An Essay / Concerning Humane Understanding / Lib. 1', has 'Intellectus A' at the bottom.

12 Locke's letters to Clarke, 17 and 31 Dec. 1686 (N.S.); I say above 'as we know it' because there may have been shorter and rougher, though also completed, drafts at an earlier date.

13 Tom. viii, pp. 49–142. ('Bibliotheque' in the title contained no accent.) The printing of the Epitome was being finished about 6 Feb. 1688 (n.s.), as indicated in Locke's letter to Clarke so dated. The title of the Epitome as printed in the Journal begins 'Extrait d'un livre anglois qui n'est pas encore publié ...'; the title of it in the separate issue made for private circulation begins 'Abrégé d'un Ouvrage ...'. (These and all other occurrences of the ellipsis '...' in the present edition are editorial.) A transcription of an extant English draft of the Epitome (Bodleian Library, MS. Locke c. 28, f. 52), and a collation of it with the French publications of the Epitome, will be included in Drafts. On the importance of the Epitome for the diffusion of Locke's doctrines on the Continent, see Gabriel Bonno, Lettres inédites de Le Clerc à Locke (Berkeley and Los Angeles, 1959), p. 5. Cf. also 10(27) below.

14 Letter to Locke, 25 Nov. 1687. In fact, the Abrégé (as distinct from the 'Extrait') carried a dedication to Pembroke, as did the Essay (see Bodleian Library, MS. Locke c. 11, ff. 195–6). The third and the future fourth Earl of Pembroke had been the dedicatees of the Shakespeare First Folio, 1623. Locke first got to know the eighth Earl of Pembroke (1656–1733) well at Montpellier in 1676 (when Pembroke was still Thomas Herbert and had not yet succeeded his brother to the Earldom). See H. R. Fox Bourne, Life of John Locke (London, 1876), I, pp. 364–5 and Maurice Cranston, John Locke: A Biography (London, 1957), pp. 166–7. He was F.R.S.; P.R.S., 1689–90.

15 Letter, 30 Dec. 1687 (n.s.); Locke apologetically explained that he could not refuse to send a copy of his treatise to the Earl of Pembroke for reasons he gives. Clarke had a completed copy (formed by instalments) of Locke's current draft of the Essay from about Feb. 1686/7; cf. Locke's letter to Clarke, 26–8 Jan. 1687 (n.s.). Tyrrell sent back to Locke in Feb. 1691/2 the draft in his possession.

16 Bodleian Library, MS. Locke b. 1, f. 109. A transcription of this and other manuscript material relating to the publication history of the Essay will be included in Drafts. A modernized excerpt from the Bassett contract is given in Cranston, op. cit., pp. 318–19. On the date, cf. also 5(19).

17 See Bodleian Library, MS. Locke f. 29, p. 36; and Tyrrell's letter cited in n. 11, above.

18 Charlotte S. Johnston, 'The printing history of the first four editions of the Essay concerning Human Understanding', Appendix II in Richard I. Aaron, op. cit., p. 315. Her sketch of the Editions could have been valuable if it had not contained so many errors.

19 Henry Ollion, Lettres inédites de John Locke à ses amis Nicolas Thoynard, Philippe Van Limborch et Edward Clarke (The Hague, 1912), pp. 191–3. Locke's letters to Limborch, 7 Aug. and 3 Dec. 1689, indicate that he had been sending a copy in proof of Bks. I–III to Le Clerc; unfortunately his letters to Le Clerc accompanying the proofs of the Essay were not preserved.

20 The actual speed at which the printers executed their task in the composition and production of the First Edition was well above the average among English presses at that time; cf. D. F. McKenzie, The Cambridge University Press, 1696–1712 (Cambridge, 1966), I, p. 103.

21 The danger of loss or fire at the printing-house may have been an additional factor.

22 Of course, it is possible that Locke always ensured that the printers had an adequate supply of his manuscript at hand. In the sixteenth and seventeenth centuries, some authors regularly attended (even resided) at the printing-house during the printing of their work; among other advantages, this arrangement secured authoritative proof-correction and prompt help for the printers when difficulties arose, e.g. when there were complicated alterations in the manuscript, the handwriting was hard to decipher, or a piece of copy was missing. My impression is that, except perhaps rarely, Locke did not visit the printing-house during the production of any edition of the Essay. For a general guide to the author-printer relationship in this period, see Percy Simpson, Proof Reading in the Sixteenth, Seventeenth and Eighteenth Centuries (London, 1935; repr. 1970).

23 P. H. Nidditch, 'A textual study of Locke's Essay: The Epistles and Book I' (privately circulated, Jan. 1962).

24 The Holt issue is listed before the Mory issue in the Catalogue of Printed Books in the British Museum, and the Mory title-page is there noted as being a cancellans. See also Wallace Nethery, 'John Locke's essay on "humane understanding"', American Book Collector, 12 (1962), 13–17.

25 See MS. cited in n. 17 above.

26 The Term Catalogues, ed. Edward Arber, II (for 1683–1696), p. 302; the entry appears under the Hilary Term (February).

27 One of the facts adduced by Johnston, op. cit., p. 313, is that 'All examined copies of the [Holt] issue have corrections, in ink', namely, 'certainly' on the last page of The Epistle Dedicatory was corrected to 'extremely', and 'some' was inserted before 'Discovery' on the first page of The Epistle to the Reader. It should be pointed out, however, that there are copies of the Holt issue which do not contain such corrections; one is in the University of Sheffield Library.

28 This is the number of copies of the Principia suggested by W. W. Rouse Ball, An Essay on Newton's 'Principia' (Cambridge, 1893), p. 67; 300–400 copies has been more recently regarded as probable by A. N. L. Munby, 'The distribution of the first edition of Newton's Principia', Notes and Records of the Royal Society of London, 10 (1952), 28–39. (750 copies were printed of the second edition (1713) of the Principia and c. 1,250 copies of the third edition (1726).) In general, whether a book produced in England in this period was written in Latin—and therefore could be read by foreigners—or in English hardly seems to have affected the number of copies of it that were printed.

29 See David Thomas's letter to Locke, 25 Apr. 1688.

30 According to Bassett's letter to Locke, 28 Feb. 1692/3, the First Edition had nearly sold out. I do not know at what price copies (in quires) were sold retail; perhaps at ten or eleven shillings; cf. McKenzie, op. cit., I, Table 16, pp. 142–3, and Norma Hodgson and Cyprian Blagden, The Notebook of Thomas Bennet and Henry Clements (Oxford, 1956); see also n. 54 below.

31 Bodleian Library, MSS. Locke c. 3, f. 160; b. 2, f. 115; b. I, f. 168: and Locke's letter to Molyneux, 20 Sept. 1692 (The Works of John Locke Esq.: In Three Volumes (London, 1714), III, p. 503; the letters (first published in 1708) thus cited here are readily found under their dates in all later editions of the Works).

32 Letter to Locke, 22 Dec. 1692 (Works, III, p. 505). All quotations (except in n. 33) from Locke's Correspondence are given in a somewhat modernized form.

33 These summaries differ from what was printed in the Second and later Editions, and run as follows (see Works, III, p. 515):

'§.28. Volition is the ordering of some Action by Thought.

§.29. Uneasiness determines the Will.

§.30. Will must be distinguished from Desire.

§.31. The greater Good in view barely consider'd determines not the Will. The Joys of Heaven are often neglected.

§.32. Desire determines the Will.

§.33. Desire is an Uneasiness.

§.34. The greatest present Uneasiness usually determines the Will, as is evident in Experience. The Reasons.

§.35. Because Uneasiness being a Part of Unhappiness, which is first to be removed in our Way to Happiness.

§.36. Because Uneasiness alone is present.

§.37. The Uneasiness of other Passions have their Share with Desire.

§.38. Happiness alone moves the Desire.

§.39. All absent Good not desired, because not necessary to our Happiness.

§.40. The greatest Uneasiness does not always determine the Will, because we can suspend the Execution of our Desires.'

34 Locke's letter to Molyneux, 23 Aug. 1693 (Works, III, p. 520); see also letter to Molyneux, 8 Mar. 1694/5 (ibid., p. 532).

35 Letter to Locke, 2 Mar. 1692/3 (Works, III, p. 512). Molyneux sent a first statement of his Problem to the author of the Epitome (n. 13), on 7 July 1688.

36 See Molyneux's letters to Locke of 23 Dec. 1693 and 17 Feb. 1693/4, and Locke's letters to Molyneux of 19 Jan. 1693/4 and 26 May 1694 (Works, III, pp. 521–4). The passage noted below, 344(4–8), n. which was first printed in the Errata of the Second Edition, stemmed from this correspondence.

37 Letter to Molyneux, 28 Mar. 1693 (Works, III, p. 513).

38 Molyneux's letter to Locke, 2 Mar. 1692/3, and Locke's letters to Molyneux, 28 Mar. 1693 and 26 May 1694 (Works, III, pp. 511, 513, 524).

39 Molyneux's letter to Locke, 2 Mar. 1692/3; Locke's letter to Molyneux, 23 Aug. 1693.

40 See further pp. l–lii below.

41 A detailed analysis of the changes that appeared at each stage of the printed Essay will be given in the Introduction to Drafts in conjunction with an analysis of the development of the Essay in draft.

42 See the postscript of Locke's letter to Clarke, following his letter to Mrs. Clarke, 9 Mar. 1688 (n.s.): 'I hear that Thomas and Susan [Stringer] are in town. I wonder I hear not a word from them. When I print my book, as I think now I shall, I would have my picture before it; therefore, pray get the picture they have of mine up to town whilst you are there, that I may take order to have a plate graved from it' (Benjamin Rand, The Correspondence of John Locke and Edward Clarke (Oxford, 1927), p. 256); and Cranston, op. cit., pp. 299, 323–4, and references cited there.

The available evidence indicates that there were two portraits of Locke by Greenhill: an oval and a larger picture, probably rectangular and about kit-cat size (i.e. less than half-length but including the hands). The oval first belonged to Dr. David Thomas; next, to Clarke, and his daughter Elizabeth's descendants; and then (by purchase in 1954) to the National Portrait Gallery, London. A reproduction of it forms the frontispiece to the present edition. The other painting was the one at issue with the Stringers; it has been lost.

43 I derive this date of publication from Locke's letter to Molyneux, 26 May 1694 (Works, III, p. 524): 'my book which is now ready and bound, and ready to be sent to you ...'.

44 I have not used copies of the Dring-Manship issue in my collation.

45 The Term Catalogues, II, p. 541; the entry appears under the Hilary Term (February), 1694/5.

46 Locke adds, in the same sentence, 'nec adhuc invenit dissertatio illa utcunque heterodoxa oppugnatorem'—'and that dissertation, heterodox as it may be, has not yet found an opponent' (Works, III, p. 612); this is curious in view of the objections that had already been brought against the Essay by John Norris and others.

47 Works, III, p. 527. Cf. 504(13 ff.) and 511(23 ff.).

48 Works, III, pp. 531, 534, 537.

49 Wynne's letter to Locke, 31 Jan. 1694/5, and Locke's reply, 8 Feb. 1694/5, were first published in Lord King, The Life of John Locke (London, 1829), pp. 189–92.

50 For further details concerning this abridgement, see Locke's letters to Molyneux, 26 Apr. 1695, 2 July 1695, and 30 Mar. 1696 ('Mr. Wynne's Abstract of my Essay is now published') (Works, III, pp. 534, 537, 546); and Wynne's letters to Locke, 20 Feb. 1694/5 and 25 June 1695 (Bodleian Library, MS. Locke c. 23, f. 113 and f. 119). Locke stated, in the first-mentioned letter, in relation to a proposed Latin translation of the Essay, that it should be abridged 'in paring off some of the superfluous repetitions, which I left in for the sake of illiterate men, and the softer sex, not used to abstract notions or reasonings'.

51 Letter to Molyneux, 10 Apr. 1697 (Works, III, p. 561). See further, n. 65 below. The draft of the Conduct was published in Posthumous Works of Mr. John Locke in 1706 and in Works, III.

52 This exemplifies a type of phenomenon of authorial revision—towards a more careful use of punctuation and a readier use of the heavier signs of punctuation—that is not only to be found in some of Locke's other works, e.g. Some Thoughts Concerning Education, but is not uncommon in English literary history; a striking case in the nineteenth century is Dickens's treatment of Oliver Twist, especially for the 1846 edition (cf. Kathleen Tillotson's edition of this novel (Oxford, 1966), pp. xxxviii–xxxix).

53 See Locke's letter to Dr. Molyneux (the brother of William Molyneux, who had died the preceding October), 25 Jan. 1698/9.

54 On date of going to press, cf. Bodleian Library, MS. Locke b. 1, f. 218, and Locke's letter to Bold, 16 May 1699 ('My Essay is going to be printed again'). On date of publication, cf. Locke's letter to Sloane, 2 Dec. 1699; the advertisement in the London Gazette (3556), 7–11 Dec. 1699; and the entry in The Term Catalogues, III, p. 176, for Hilary Term (February), 1699/1700 (this entry is not indexed in the Title-Index, but only in the Author Index).

Most advertisements of books in this period did not mention prices; the advertisement in the London Gazette above-mentioned is no exception. According to the Bibliotheca Annua, 1699, copies of the Fourth Edition of the Essay were for sale at 14s. retail; cf. Hodgson and Blagden, op. cit., p. 158.

55 The surviving drafts of intended additions in Bodleian Library, MS. Locke e. 1, have contents that are in general closely similar to, but not identical with, the published additions: that manuscript was a source of, but did not itself serve as, printers' copy.

56 These first four mispaginations occurred also in the Second and Third Editions.

57 Bodleian Library, MS. Locke b. 5, No. 14; 'small' is added interlinearly. The reference here to the '31 chapter' was a slip (not necessarily Locke's) for '21 chapter'; cf. the new final paragraph of The Epistle to the Reader,14(15–18) below. See further, on the Locke-Stillingfleet material in the Fifth Edition, p. xiii below.

That Locke had already before this codicil was written finally ceased to work at further revision of the Essay is suggested by Lady Masham's letter to Leibniz, 3 June 1704 (reproduced in C. J. Gerhardt, Die philosophischen Schriften von Gottfried Wilhelm Leibniz, III (Berlin, 1886), p. 351), in which, speaking for Locke (who might, however, have been only making an excuse for not engaging in epistolary discussion with Leibniz), she states: 'His want of health he says now, and the little remains he counts he has of life, has put an end, to his enquiries into philosophical speculations.'

58 See 12(5–6), n., below.

59 On the date of publication of Coste's translation in 1700, cf. Limborch's letter to Locke, 20 July 1700 (N.S.): 'Ante hebdomadas aliquot tuo nomine mihi datus est praestantissimus tuus de Intellectu Humano liber in linguam Gallicani versus'—'I was given some weeks ago on your behalf a copy of your distinguished book Concerning Human Understanding in a French translation' (Works, III, p. 644). Coste gives a considerable number of corrections and additions to his first edition in his letter to Leibniz, 25 Aug. 1707 (reproduced in Gerhardt, op. cit, pp. 392–9). On Coste, see article on him in Dictionary of National Biography, XII, pp. 275–6. He edited or translated the works of many authors; his publications included a learned edition of Montaigne's Essais and a French translation of Newton's Opticks.

60 On the history of translating the Essay into Latin, see Fox Bourne, op. cit., II, p. 273, Bonno, op. cit., pp. 38 and 40, and the letters between Locke and Molyneux in Works, III, pp. 530–1, 533, 535–6, 538–9, 541–2, 544–6, 548–9, 551, 557, 566, 571, 573, and 585; also Ollion, op. cit., pp. 191, 197, 198–203.

61 The exception of correspondence is made because of the separate publication, under the editorship of E. S. de Beer, of all available letters (or drafts of letters) from or to Locke. But excerpts from, or summaries of, relevant letters will be included in Drafts.

62 Or till the publication of the Fifth Edition in 1706 in so far as there are manuscripts by other persons relating to the history of this Edition.

63 I call Draft α the very early draft of the Essay that is in the Public Record Office (P.R.O. 30/24/47/7).

64 See p. 14 and n. 13 above.

65 Locke's willingness to have this posthumously published is expressed in his letter to Peter King, 4 Oct. 1704, excerpts from which are printed in Cranston, op. cit., pp. 477–8. The manuscript of the Conduct is Bodleian Library, MS. Locke e. 1, chiefly pages numbered 62–260; on p. 62 of it is written, inter alia: 'B:IV C: XX / Of the Conduct of the under / standing'. Locke originally thought of adding his account of the topic, thus designated, as a new chapter that was to be numbered XX in the Fourth Edition; instead thus numbered there is the chapter entitled 'Of wrong Assent, or Errour', which is numbered XIX in the first three Editions. Cf. p. xxviii above.

66 Although copy for the first edition of books of this period has usually disappeared, there are some survivals; for example, Book I of Milton's Paradise Lost, and several works of Newton's, including the Principia, Opticks, and Arithmetica Universalis.

67 A photo-reproduction of the Holt issue using the copy in the University of Cambridge Library has been published by the Scolar Press (Menston, Yorkshire, 1970); this also contains, in an appendix, the frontispiece, title-page, and some of the alterations to the text of the First Edition that belong to the Churchill-Manship issue of the Second Edition.

68 See p. xxxii above.

69 Another contemporary illustration of the care of an author (or his deputy) about the printing and proofs of his book is given by the history of the first edition of Newton's Principia. Some of the evidence is published in The Correspondence of Isaac Newton, II (1676–1687) (Cambridge, 1960). In Halley's letter to Newton, 7 June 1686, he says: 'I here send you a proof of the first sheet of your Book, which we think to print on this paper, and in this Character; if you have any objection, it shall be altered: and if you approve it, wee will proceed ... The Printer begs your excuse of the Diphthongs, which are of a Character a little bigger, but he has some a casting of the just size. This sheet being a proof is not so clear as it ought to be; but the letter is new, and I have seen a book of a very fair character, which was the last thing printed from this set of letter; so that I hope the Edition may in that particular be to your satisfaction.' See further I. B. Cohen's book cited in n. 83 below.

70 They are rather sparse in Locke's extant manuscripts, e.g. Bodleian Library, MS. Locke, e. I. His annotations of The Reasonableness of Christianity (Houghton Library, Harvard University) give direct evidence of an interest in spelling and the distinction between capital and small letters for his publications.

71 Specially relevant are: the copy of Locke's Two Treatises of Government at Christ's College, Cambridge, with its numerous annotations by Locke or Coste (see Peter Laslett, John Locke: Two Treatises of Government (Cambridge, 1960; second edition (with amendments), 1970), the 1690–1713 editions of that work, and the 1693–1705 editions of Locke's Some Thoughts Concerning Education.

72 So far as I have discerned, the copies of the Fourth Edition are unlike in matters covered by the present edition only by very rare differences in the presence or absence of punctuation, which result from accidents, and the wear-and-tear, of the printing(-off) process. In that Edition, and the other Editions, of the Essay intentionally different states occur only rarely (see 5(16), n., 6(14), n., and p. 817).

73 A classical expression of this standpoint on textual editing is A. E. Housman, 'The application of thought to textual criticism', Proceedings of the Classical Association, XVIII (1921), 67–84. See also Greg's paper cited in the next note.

74 My treatment of the copy-text generally conforms with the text-critical principles propounded by W. W. Greg, 'The rationale of copy-text', Studies in Bibliography, III (1950–1), 19–36 (reprinted in Greg's Collected Papers, ed. J. C. Maxwell (Oxford, 1966), pp. 374–91); see also his The Editorial Problem in Shakespeare, third edition (Oxford, 1954), especially pp. vii–lv. I have not adopted Greg's terminology of 'substantives' and 'accidentals' (the distinction being between those readings 'that affect the author's meaning or the essence of his expression, and others, such in general as spelling, punctuation, word-division, and the like, affecting mainly its formal presentation') which can occasion awkward ambiguities and carries no convenient adjectival and adverbial cognates with it.

75 It is hoped that an edition of the Locke-Stillingfleet materials will be published in due course to complement this volume and Drafts: the whole of Locke's extant writings for and defences of the Essay will then be available in critical editions.

76 Cf. p. xxx above.

77 Conventionally, the hyphen does double duty, both as a link between unamalgamated words and as a symbol, at the end of a line, to designate continuity within a single word. It is possible that, in a few instances, I have mistakenly interpreted a hyphen in the copy-text as being applied in one use instead of the other, and that, in some cases, a regrettable ambiguity faces the reader of the Text by the presence of a hyphen at the end of a line (but the register elucidates these).

78 The copies of the early printed versions that I have examined at length are the ones in the following libraries or ownership. A number of other copies of each listed item have also been consulted.

1: University of Bristol Library, University of Cambridge Library, University of Sheffield Library, British Museum.

2: My own copy, University of Bristol Library, University of Sheffield Library.

3: City of Bristol Public (Reference) Library.

4: My own copy, City of Bristol Public (Reference) Library, University of Hull Library.

5: My own copy, University of Bristol Library.

6: British Museum.

W: University of Sheffield Library.

Coste: My own copy, University of Birmingham Library.

Coste2: University of Leeds Library.

79 Cf. p. xxxv above.

80 To save duplicating references to Coste's departures from the sectional summaries in the copy-text, such departures are annotated only in the critical apparatus for the main Text, and not for the Table of Contents.

81 P. H. Nidditch, A Bibliographical and Text-Historical Study of the Early Printings of John Locke's 'Some Thoughts Concerning Education' (Sheffield, 1972), p. 5.

82 In this context, 'printers' means the compositors and correctors employed in a printing-house.

83 See Helen. Darbishire (ed.), The Manuscript of Milton's Paradise Lost, Book I (Oxford, 1931), and Vol. II of Fletcher's work cited in n. 87 below.

The first edition of the Principia generally follows with complete fidelity the formal features of the manuscript (in the Library of the Royal Society) used by the printer; the departures which occurred intentionally (as distinct from accidental misprintings) consist of, for example, a semicolon occasionally replacing a comma, and a systematic preference of '&' for 'et'. Unfortunately, matters of this sort are not discussed in I. B. Cohen, Introduction to Newton's 'Principia' (Cambridge, 1971); and in Koyré and Cohen's edition of the Principia 'the variants in the Apparatus Criticus do not include changes in spelling, capitalization, punctuation, or italicization' (Cohen, op. cit., p. 34, n. 2), these changes therefore remaining unannotated since that edition has no separate register of formal variants. On the other hand, the printers of Newton's Opticks (University of Cambridge Library, Add. 3970. 3) punctuated, capitalized, etc., according to their own preferred style; e.g. they much more frequently capitalized nouns and adjectives, and much more densely punctuated (but they nearly always preserved the original full stops and colons). There is no critical-text edition of the Opticks.

84 On all this, see Simpson, op. cit., e.g. pp. 51–6, 71–8, 86, 95, 103–4; Mindele Treip, Milton's Punctuation and Changing English Usage, 1582–1676 (London, 1970); McKenzie, op. cit. I, p. 118; and Joseph Moxon, Mechanick Exercises on the Whole Art of Printing (London, 1683).

85 The additions to the Errata page in Tyrrell's copy of the First Edition are neatly and attractively written; the script is entirely different from Tyrrell's flowing hand.

86 Laslett, op. cit., pp. 9–10. In round figures, the collation in Laslett's edition of the Two Treatises of Government contains 2,000 entries, three-quarters of these recording formal variants; however, many formal variants of various kinds are not included (cf. ibid., p. 447).

87 Cf., e.g., C. H. Herford and P. and E. Simpson's edition of The Works of Ben Jonson (Oxford, 1925–52) and H. F. Fletcher's edition of John Milton's Complete Poetical Works (Urbana, 1943–8).

88 The extensive presence of the annotations of formal variants could be a large and thorny obstacle in the way of a reader's keeping track of the annotations of material variants; hence the unusual arrangement has been adopted under which the bulk of the formal variants have been collectively registered separately from the material variants.

89 Locke explained this device in his 'A New Method of a Common-Place-Book' (Works, III, p. 488); this work was originally published in French in Tom. ii of Le Clerc's Bibliotheque. See also John Harrison and Peter Laslett, The Library of John Locke, 2nd ed. (Oxford, 1971), pp. 33–4.

90 The Introduction to Drafts will include some account of Locke's language. There is at present no detailed modern study of the grammar or vocabulary of an English prose writer of Locke's period; among books useful as introductions or for comparative purposes are A. C. Partridge, Tudor to Augustan English (London, 1969) and works cited in its Bibliography (pp. 226–32), R. D. Emma, Milton's Grammar (The Hague, 1964), Robert Adolph, The Rise of Modern Prose Style (Cambridge, Mass., 1968), Stanley E. Fish, ed., Seventeenth-Century Prose: Modern Essays in Criticism (New York, 1971), R. F. Jones, et al., The Seventeenth Century (Stanford, 1951), pp. 10–160, Brian Vickers, Francis Bacon and Renaissance Prose (Cambridge, 1968), George Williamson, The Senecan Amble. A Study in Prose Form from Bacon to Collier (London, 1951); see also Rosalie Colie, 'The essayist in his Essay' in John W. Yolton, ed., John Locke: Problems and Perspectives (Cambridge, 1969), pp. 234–61.