time of Bagehot's death, is in a little the worst condition of all; but that on Mary Wortley Montagu is not much better. Nor is this by any means confined to extracts: the foot-notes will show more than one case where the sense of his own writing is destroyed by a misread word whose correct reading is easy to guess. In this list belong also a mass of mere typographical errors and slips of the pen (as George III. for George I., Queen Anne for George II., etc.), sometimes of a most annoying kind; for instance, in two cases a misprinted date caused a long search for a quotation in the wrong quarter.

Another remarkable and curiously balanced sort consists in the misplacement of quotation marks, either crediting Bagehot with the writing of others or vice versa; there are, I think, just the same number of each. For instance, a half-page of Lady Louisa Stuart's is printed as his own in the article on Mary Wortley Montagu; a half-page of Lord Mahon in that on William Pitt; several lines of Lord Macaulay in "Lombard Street" (mangled as usual); and there are one or two more. On the other hand, he partially requites Lady Louisa with a couple of lines, and gives some to Le Marchant in the sketch of Lord Althorp and to Catlin in "Economic Studies," and there are other instances.

Again, the number of cases where the sense is exactly inverted by the misplacement of negatives or the reversal of the place of alternatives in a sentence is something incredible; I have kept no exact account, but there must be well toward twenty in the five volumes. A few of these have been changed outright (and noted in the table of alterations); but in general, attention has been called to them in foot-notes.

The department of distorted quotations is recruited from so many different sources of error that it is hard to know where to begin. Of course inevitable lapses of memory are the chief cause for the corrupt state of minor quotations and anecdotes. No miscellaneous writer can possibly go back to all his original sources to verify his "points," it would take a lifetime to write a volume in any such way. The maxim of the proof-room is, that quotations are always to be assumed as wrong; and it is curious how seldom the fact is otherwise. And very likely Bagehot's memory seems to me to have worked with more than usual crookedness because I have not had occasion to explore the maze of any other man's; but it seems impossible that any one else can ever have remembered so many non-existent things, or so often wrongly accredited his quotations or introduced them with total irrelevance to the context. The list of these is long, but I will not

even attempt to cite them. I should perhaps add to this list a favorite performance of his which is quite as exasperating to an editor as the worst of the others, citing passages from a first edition which were expunged in later ones, but without hinting at such effacement; this trick in the cases of Mackintosh, Montalembert, and Gibbon, cost several days of wasted time.

But the abominably corrupt state of his longer quotations — some of which are simply miracles of mangling, and much the greater part of which are more or less misquoted — cannot be accounted for by bad memory; for it can hardly be supposed that he trusted solely to his memory for whole pages. I can only guess at the modus operandi: but my guess is, that he copied his extracts by scribbling off the catch-words, trusting to his memory to fill out the skeleton when he prepared his copy for the press, and never comparing the outline with the original (thus often leaving out words, phrases, and occasionally even whole lines); that as might be expected, his memory was treacherous when the time came to write out his extracts in full; that (which is certain) the printers made mistakes in his copy, and he did not correct them; and as a result (which is most certain of all), that his quoted matter cannot be matched in the language for (sometimes absurd) divergence from the original.

It would seem at first thought that there would be no trouble in dealing with these things, that the only thing necessary was to restore the true text, if it could be found, and there all difficulty ended; but in fact it has not been at all easy to decide in every case what to do, and I am by no means sure I have invariably made the best decision. Many of the alleged quotations could not be found at all, even when every attainable scrap of an author's published works or words was at hand. In some of these cases I can definitely prove that they have no existence, and sometimes can even show what he manufactured his "quotation" out of; but in far the greater number (and a list of the undiscoverable things which seem to lie easily at hand, and which I supposed to do so till after thorough and fruitless search, would excite surprise) I can only conjecture that he has credited them to the wrong authors, and have either passed them silently or transferred the problem to the reader by a foot-note. Sometimes it is a "made-up " quotation, -a fiction founded on fact, so to speak (there are instances of this in the articles on Sterne and Pitt, and elsewhere); sometimes the "quotation " gives the general sense of the original, but in a totally different form of words in both these cases foot-notes are the obvious propriety. But between the latter and the ones so

slightly blundered as to involve only silent correction, there lie every grade of mangling, a border-land where judgment is difficult, and I have sometimes substituted a correct form where another might have left the corrupt form standing and annotated it; I can only say that in no case of the sort have I tampered with Bagehot's own words, and in nearly every one I have called attention to the great difference of the correct quotation from Bagehot's text.

That the source of every quotation has been given whenever possible, follows of course. Apart from the question of accuracy, a reader of any author has a right to know this, in order to examine the context and follow the author in his track of reading; and judging from my own experience, no other service in editing save the explanation of obscure allusions is comparable to this, nor any lack an equal hindrance and exasperation. For the same reason, I have indicated where it was feasible the main sources whence he drew the facts for his review articles, and that of many special biographic or historical items. My object has been, to make the volumes as handily useful to the least scholarly reader as might be; I have assumed that many would be glad to use the articles as a base for some further reading if it was made easy, who could not or would not engage in any research requiring much time.

It would be ungrateful not to mention here the two assistants who have lightened my work and added to its value in its later stages. During the last few months of 1888 Miss T. J. O'Connell did much capable and faithful searching at the British Museum; but my chief debt is to Miss Fanny G. Darrow of Boston, who ransacked libraries in behalf of the work for a year, and without whose zeal, patience, and book-lover's "nose" for the probable place of extracts, the measure of its merits would have fallen far below what it is.

I come next to a most delicate subject, on which I have risked much vituperation; namely, my dealings with the murdered grammar and impossible syntax with which all Bagehot's writings abound. No writer of eminence in modern times (so far as I know) has treated so defiantly the primary grammatical rules of the English language, or the first principles of construction in any language. He was a business man, and he is an adept at "business talk" as frequently heard among that class of men, fectly lucid as to matter and perfectly incoherent as to structure, utterances which no man can mistake and no man can parse. There are sentences in his works which are no more English than they are Chinese, and yet are not in the least indistinct as to

-per

meaning; indeed, Mr. Giffen says he sometimes wrote bad grammar purposely to make his meaning clearer, which is a startling proposition. He must have known the difference between a principal and a subordinate clause, but he put the knowledge to no practical use. His moods are kept pretty well in hand; but his tenses are at the mercy of fate and chance, and some of his paragraphs are perfect see-saws of past and present, mixed with the wildest indifference not only to grammar but to sense. His verbs have no certainty of agreeing with his nouns in number and person; his personal pronouns are as defiant of the trammels of singular and plural relations as his verbs are of the fetters of tense; and his relatives may or may not refer to the noun they follow.

That no editor has any business to rewrite a line or change a substantive word of his author's text is self-evident; and that the substitution of any language of mine for that of Walter Bagehot would be the summit of impertinence and presumptuous folly is equally evident. What readers wish to know and have a right to know is, what Bagehot said, not what his editor thinks he ought to have said. Therefore, in no case have I meddled with the structure of a sentence in any way; in a few cases I have called attention to the entanglement of the syntax, but I have not even attempted to mend such atrocities as "The period at which the likeness was attempted to be taken" (beginning of the "English Constitution"), or other like gems of English. But I do not think even editorial fidelity or reverence for the memory of a great man (and I cannot better gauge my own for Walter Bagehot than by saying that I believe this edition is a higher service to the public than any original work I could do) binds me to allow a plural noun to remain coupled with a singular verb (or vice versa), or a singular pronoun in one clause set off against a plural one in the following like clause, or a present and a past tense similarly yoked together in a most discordant union, merely because the great man did not read his proofs and a patent slip of the pen remained uncorrected. I do not believe even he, little as he cared for such things, would wish to have all the rags and tatters of his haste and slovenliness scrupulously saved up and exhibited to posterity, any more than a public speaker would care to have a phonograph record an accidental hiccough; nor do I believe that even the most devoted admirers of Bagehot, to whom every word is worth preserving as instinct with the flavor of that rich mind (among whom I count myself), care to have their senses jarred upon by such purely accidental slips. Nevertheless, I recognize the right of the public to know just what their author wrote and how he

left his text; that he wrote carelessly and did not read his proofs is in itself an item of interest in comprehending him. And still more, I owe both to them and to myself to give the minutest information just how far I have tampered with the text, so that they may not fear that they are reading a mangled and wantonly altered version, and I may not be suspected of meddling with his language. I have therefore kept a scrupulous account of all the changes, even the minutest, (except such as are made by the insertion of words or letters, in which case the additions are invariably put in brackets, - or by foot-notes,) and give them in a separate table. By this means, any one who finds comfort in knowing how badly his author could write can do so, and where no notice is given may be sure he is reading Bagehot undefiled.

That all extracts in foreign languages are translated, ought to be more a matter of course than it is: in anything designed for wide popular reading, neglect to do so is either laziness or swagger. The object being that all readers shall have the fullest understanding and enjoyment with the least friction, it is absurd to lock up any portion out of the reach of four-fifths of them; and it is not the business either of a writer or an editor to impose penalties for defective education. There is of course one palpable exception to this, — where an extract is cited as a sample of style instead of matter; which in general excludes translation of all poetry as well as of some prose. But curiously enough, not a single quotation of Bagehot's from any foreign author is given to illustrate style: even the verses from Sophocles in the essay on Shelley are cited only as an instance of classic bareness of decoration, and he quotes poems from Béranger only to illustrate that poet's philosophy of life. The worst translation possible, therefore, would be better than none; while in fact Mr. Walter Learned has graced this edition with several excellent translations of Béranger (some of which I think much the finest of any yet executed), and for the others I have taken the best I could find.

The foot-notes marked "B." are Bagehot's; those of Mr. Hutton are marked "R. H. H."; my own are signed "Ed." The latter is only added, however, to controversial or corrective notes; simple references to sources of quotations are left uncredited, though all but a very few are new to this edition, and some of the very few in previous ones are either wrong or unintelligible. By the latter I mean page references, which are the most exasperating of traps, since one is never sure of having the same edition as that cited, and the page number simply confuses him on any other. For this reason I have avoided them rigorously, and made references to