How to Make an Index

The two points made by the "Scholar" were: (1) That the making of a general subject index to the catalogue proposed by the authorities of the British Museum would be a waste of money; (2) That it was a great evil for the five-yearly indexes originated by Mr. Fortescue to be discontinued.

Now let us see what is to be said with authority on these points.

Mr. Fortescue said:

"Last Autumn … I read with respectful astonishment a letter to 'The Times' from a writer who preferred to veil his identity under the modest signature of 'a Scholar.' There I read that 'the studious public of this country and Europe in general have been surprised by the news that the authorities of the British Museum seriously contemplate the compilation of a subject index to the vast collection of printed books in that library.' I can assure you that the surprise of the studious public and of Europe in general cannot have surpassed my own when I thus learned of what the authorities were seriously contemplating. Nevertheless, it left me able, I thought, to discern that their vast conceptions had not been so fortunate as to gain the approval of 'a Scholar' and to marvel whence The Times and other great journals had drawn their truly surprising information. Some of the arguments put forth in sundry criticisms of the 'scheme' showed how much thought had been bestowed upon matters which then first dazzled my bewildered imagination. It may come some day (who shall say what will not?), this General Index, or it may never come. But up to the present moment I am aware of no authority who is seriously contemplating so large a venture unless perhaps it be 'a Scholar' himself."

Then as to the five-yearly indexes Mr. Fortescue said:

"Experience has taught us that there is no form of subject-index which the public values so highly as one which gives the most recent literature on every possible subject. And to meet this manifest want we shall certainly continue to issue, with all the latest improvements I hope, the modest Indexes which we have hitherto published in five-yearly (I am afraid as President of The Library Association I should say 'in quinquennial') volumes. The Museum sweeps its net so wide and in such remote seas that a more or less complete collection of books on almost every subject or historical event is gathered within it for future students. To take only two incidents from the last year or two, the next index will contain not less than a hundred and forty books and pamphlets, in almost every European tongue, on the Dreyfus case, and from four to five hundred books on the present war in South Africa. Such bibliographical tests have more than an ephemeral or immediate value. They will remain as records of events or phases of thought long after their causes shall have faded from all but the page of history."

Of late years the dictionary catalogue has come very largely into use in public libraries. This consists of a union of catalogue of authors and index of subjects which is found to be very useful and illuminating to the readers in free libraries, most of whom are probably not versed in the niceties of bibliographical arrangement, but are more likely to want a book on a particular subject than to require a special book which they know. Mr. Cutter has written the history of the dictionary catalogue in the United States Special Report (pp. 533-539), and he traces it back in America to about the year 1815.

Excellent specimens of these dictionary catalogues have been produced. They are of great value to the ordinary reader at a small public library, but I venture to think that to construct one for a large library is a waste of power, because if several large libraries of a similar character do the same thing, there is constant repetition and considerable loss by the unnecessary outlay. If a fairly complete standard index were made, it could be used by all the libraries, and in return the libraries might unite to pay its cost. I am pleased to know that Mr. Fortescue prefers to keep index and catalogue distinct. He said in his address:

"I have formed, so far as I know, but one dogmatic conviction, and it is this: that the best catalogue which the art of man can invent is a catalogue in two inter-dependent yet independent parts; the first and greater part an alphabetical catalogue of authors, the second and lesser part a subject-index. I know well that I shall be told that I am out of date, that such an opinion is as the voice of one crying in the wilderness—that the dictionary catalogue has won its battle—but even so, perhaps the more so, do I feel it the part of a serious and immovable conviction to declare my belief that—for student and librarian alike—this twofold catalogue, author and subject each in its own division, is the best catalogue a library can have, and that the dictionary catalogue is the very worst. But whatever may be our individual opinion on this head, it is only necessary to enter into a very simple calculation to see that if the dictionary system could have governed the rules of the British Museum Catalogue it would by now have consisted of not less than twelve million entries; and assuredly it would have been neither completed nor printed to-day."

CHAPTER VI.

General Rules for Alphabetical Indexes

"In order to guard against blunders Bayle proposed that certain directions should be drawn up for the guidance of the compilers of indexes."

THESE rules, originally drawn up by a committee of the Index Society, were primarily intended for the use of indexers making indexes of indexless books to be published by the society, which, being produced separately from the books themselves, needed some introductory note. In all cases, however, some explanation of the mode of compilation should be attached to the index. The compiler comes fresh from his difficulties and the expedients he has devised to overcome them, and it is therefore well for him to explain to the user of the index what those special difficulties are.

The object of the Index Society was to set up a standard of uniformity in the compilation of the indexes published by them. Although rigid uniformity is not needed in all indexes, it is well that these should be made in accordance with the best experience of past workers rather than on a system which varies with the mood of the compiler. It is hoped that the following rules may be of some practical use to future indexers.

In the eighth chapter of How to Catalogue a Library there are a series of rules for making a catalogue of a small library in which are codified the different points which had been discussed in the previous chapters. In the present chapter the Index Society rules are printed in italic, and to them are now added some illustrative remarks. There is necessarily a certain likeness between rules for indexing and rules for cataloguing, but the differences are perhaps more marked. At all events, the rules for one class of work will not always be suitable for the other class.

1. Every work should have one index to the whole set, and not an index to each volume.

An index to each volume of a set is convenient if a general amalgamated index to the whole set is given as well; but a work with several indexes and no general one is most inconvenient and irritating, while to have both seems extravagant. If, however, the author or publisher is willing to present both, it is not for the user of the book to complain.

2. Indexes to be arranged in alphabetical order, proper names and subjects being united in one alphabet. An introduction containing some indication of the classification of the contents of the book indexed to be prefixed.

In an alphabetical index the alphabet must be all in all. When the alphabet is used, it must be used throughout. There is no advantage in dividing proper names from subjects, as is so often done, particularly in foreign indexes. Another objectionable practice frequently adopted in the indexes of periodical publications is to keep together the entries under the separate headings used in the journal itself, and thus to have a number of distinct alphabets under different headings. This union of alphabetical and classified indexing has been condemned on a former page, and need not here be referred to further.

In the case of large headings the items should be arranged in alphabetical order under them. There is occasionally a difficulty in carrying this out completely, but it should be attempted. We want as little classification as possible in an alphabetical index. Mr. W. F. Poole wisely said in reference to the proposal of one of his helpers on the Index of Periodical Literature to place Wealth, Finance, and Population under the heading of Political Economy: "The fatal defect of every classified arrangement is that nobody understands it except the person who made it and he is often in doubt."

3. The entries to be arranged according to the order of the English alphabet. I and J and U and V to be kept distinct.

There are few things more irritating than to find the alphabet confused by the union of the vowel i with the consonant j, or the vowel u with the consonant v. No doubt they were not distinguished some centuries ago, but this is no reason why they should again be confused now that they are usually distinct. There may be special reasons why they should be mixed together in the British Museum Catalogue, but it is not evident that these are sufficient.

The only safe rule is to use the English alphabet as it is to-day in an English index. One of the rules of the American Library Association is: "The German ae, oe, ue always to be written ä, ö, ü, and arranged as a, o, u." By this Goethe would have to be written Göthe, which is now an unusual form, and I think it would be better to insist that where both forms are used, one or other should be chosen and all instances spelt alike. It is a very common practice to arrange ä, ö, ü, as if they were written ae, oe, ue; but this leads to the greatest confusion, and no notice should be taken of letters that are merely to be understood.

4. Headings consisting of two or more distinct words are not to be treated as integral portions of one word; thus the arrangement should be:

The perfect alphabetical arrangement is often ignored, and it is not always easy to decide as to what is the best order; but the above rule seems to put the matter pretty clearly. If no system is adhered to, it becomes very difficult to steer a course through the confusion. When such entries are printed, a very incongruous appearance often results from the use of a line to indicate repetition when a word similar in spelling, but not really the same word, occurs; thus, in the above, Grave surname, Grave substantive, and Grave adjective must all be repeated. It is inattention to this obvious fact that has caused such ludicrous blunders as the following:

"Mill on Liberty

–– on the Floss."18

"Cotton, Sir Willoughby,

–—, price of."

"Old age

–– Artillery Yard

–– Bailey."

These are all genuine entries taken from books, and similar blunders are not uncommon even in fairly good indexes; thus, in the Calendar of Treasury Papers, 1714-1719, issued by the Public Record Office, under Ireland are the following entries:

"Ireland, Mrs. Jane, Sempstress and Starcher to King William; cxcvii. 32.

… Attorney General of, See Attorney General, Ireland."

Then follow nearly two columns on Ireland with the marks of repetition (…) throughout.

The names of streets in the Post Office Directory are now arranged in a strict alphabetical order on the lines laid down in this rule; thus we have:

"White Street

White's Row

White Heart

Whitechapel."

Again:

"Abbott Road

Abbott Street

Abbott's Road."

Again:

"King Square

King Street

King and Queen Street

King David Street

King Edward Road

King William Street

King's Arms Court

King's Road

Kinglake Street

Kingsbury Road

Kingsgate Street."

Sometimes there is a slip, as might be expected in so complicated a list of names. Thus in the foregoing sequence Kinghorn Street comes between King William Street and King's Arms Court, while I think it ought to come immediately before Kinglake Street; but, after all, this is a matter of opinion. Strattondale Street comes before Stratton Street; but this is merely a case of missorting.

There is one piece of alphabetisation which the editor of the Post Office Directory has always adopted, and that is to place Upper and Lower under those adjectives, and Old Bond Street under Old, and New Bond Street under New. These two names belong to what is practically one street (although each division is separately numbered), which is always spoken of as Bond Street, and therefore for which the majority of persons will look under Bond. South Molton Street is correctly placed under South because there is no North Molton Street, and the street is named after South Molton; while South Eaton Place is merely a continuation of Eaton Place. Some persons, however, think that names should be treated as they stand, and that we should not go behind them to find out what they mean.

5. Proper Names of foreigners to be arranged alphabetically under the prefixes—

but not under the prefixes—

It is an acknowledged principle that when the prefix is a preposition it is to be rejected; but when an article, it is to be retained. When, however, as in the case of the French Du, Des, the two are joined, it is necessary to retain the preposition. This also applies to the case of the Italian Della, which is often rejected by cataloguers. English names are, however, to be arranged under the prefixes:

because these prefixes are meaningless in English, and form an integral part of the name.

Whatever rule is adopted, some difficulty will be found in carrying it out: for instance, if we consider Van Dyck as a foreigner, his name will appear as Dyck (Van); but if as an Englishman, his name will be treated as Vandyck.

A prefix which is translated into the relative term in a foreign language cannot be considered as a fixed portion of the name. Thus Alexander von Humboldt, when away from his native Germany, translated his name into Alexandre de Humboldt. The reason why prefixes are retained in English names is because they have no meaning in themselves, and cannot be translated. There is a difficulty here in respect to certain names with De before them; for instance, the Rothschilds call themselves De Rothschild, but when the head of the family in England was made a peer of the United Kingdom he became Lord Rothschild without the De. In fact, we have to come to the conclusion that when men think of making changes in their names they pay very little attention to the difficulties they are forging for the cataloguer and the indexer.

In this rule no mention is made of such out-of-the-way forms as Im Thurn and Ten Brink. It is very difficult to decide upon the alphabetical position of these names. If the indexer had to deal with a number of these curious prefixes, it would probably be well to ignore them; but when in the case of an English index they rarely occur, it will probably be better to put Im Thurn under I and Ten Brink under T.

With respect to the translation of foreign titles, the historian Freeman made a curious statement which is quoted in one of the American Q.P. indexes. Freeman wrote:

"No man was ever so clear [as Macaulay] from the vice of thrusting in foreign words into an English sentence. One sees this in such small matters as the accurate way in which he uses foreign titles. He speaks, for instance, of the 'Duke of Maine,' the 'Count of Avaux,' while in other writers one sees the vulgarism of the Court Circular, 'Duke de Maine,' 'Duc de Maine,'—perhaps 'Duc of Maine.'"

Duke de Maine and Duc of Maine may be vulgar, they are certainly incorrect; but I fail to see how it can be vulgar to call a man by his right name—"Duc de Maine." I do not venture to censure Macaulay, but for lesser men it is certainly a great mistake to translate the names of foreigners, in spite of Freeman's expression of his strong opinion.

6. Proper names with the prefix St., as St. Albans, St. John, to be arranged in the alphabet as if written in full—Saint. When the word Saint represents a ceremonial title, as in the case of St. Alban, St. Giles, and St. Augustine, these names are to be arranged under the letters A and G respectively; but the places St. Albans, St. Giles's, and St. Augustine's will be found under the prefix Saint. The prefixes M' and Mc to be arranged as if written in full—Mac.

This rule is very frequently neglected, more particularly in respect to the neglect of the difference between Saint Alban the man and St. Albans the place.

7. Peers to be arranged under their titles, by which alone in most cases they are known, and not under their family names, except in such a case as Horace Walpole, who is almost unknown by his title of Earl of Orford, which came to him late in life. Bishops, deans, etc., to be always under their family names.

About this rule there is great difference of opinion. The British Museum practice is to catalogue peers under their surnames, and the same plan has been adopted in the Dictionary of National Biography. It is rather difficult to understand how this practice has come into being. There are difficulties on both sides; but the great majority of peers are, I believe, known solely by their titles, and when these noblemen are entered under their family names cross references are required because very few persons know the family names of peers. The Library Association and Bodleian rules adopt the common-sense plan of entering noblemen under their titles, and Mr. Cutter gives some excellent reasons for doing this, although he cannot make up his mind to run counter to a supposed well-established rule. Mr. Cutter writes:

"Stanhope Philip Dormer, 4th Earl of Chesterfield.... This is the British Museum rule and Mr. Jewett's. Mr. Perkins prefers entry under titles for British noblemen also, in which I should agree with him if the opposite practice were not so well established. The reasons for entry under the title are that British noblemen are always spoken of, always sign by their titles only, and seldom put the family name upon the title-pages of their books, so that ninety-nine in a hundred readers must look under the title first. The reasons against it are that the founders of noble families are often as well known—sometimes even better—by their family name as by their titles (as Charles Jenkinson, afterwards Lord Liverpool; Sir Robert Walpole, afterwards Earl of Orford); that the same man bears different titles in different parts of his life (thus P. Stanhope published his History of England from the Peace of Utrecht as Lord Mahon, and his Reign of Queen Anne as Earl Stanhope); that it separates members of the same family (Lord Chancellor Eldon would be under Eldon, and his father and all his brothers and sisters under the family name, Scott), [Mr. Cutter forgot that Lord Eldon's elder brother William was also a peer—Lord Stowell] and brings together members of different families (thus the earldom of Bath has been held by members of the families of Chandé, Bourchier, Granville and Pulteney, and the family name of the present Marquis of Bath is Thynne), which last argument would be more to the point in planning a family history."

The advocates of the practice of arranging peers under their family names make much of the difficulties attendant on such changes of name as Francis Bacon, Viscount St. Alban's, Benjamin Disraeli (afterwards Earl of Beaconsfield), Sir John Lubbock (now Lord Avebury), and Richard Monckton Milnes (afterwards Lord Houghton). These, doubtless, are difficulties, but I believe that they amount in all to very few as compared with the cases on the other side.

This is a matter that might be settled by calculation, and it would be well worth while to settle it. Mr. Cutter says that ninety-nine in a hundred must look under the title first, but I doubt if the percentage be quite as high as this. If it were, it ought to be conclusive against any other arrangement than that under titles.

Moreover, these instances do not really meet the case, for they belong to another class, which has to be dealt with in cataloguing—that is, those who change their names. When a man succeeds to a peerage he changes his name just as a Commoner may change his name in order to succeed to a certain property.

8. Foreign compound names to be arranged under the first name, as Lacaze Duthiers. English compound names under the last, except in such cases as Royston-Pigott, where the first name is a true surname. The first name in a foreign compound is, as a rule, the surname; but the first name in an English compound is usually a mere Christian name.

This rule is open to some special difficulties. It can be followed with safety in respect to foreign names, but special knowledge is required in respect to English names. Of late years a large number of persons have taken a fancy to bring into prominence their last Christian name when it is obtained from a surname. They then hyphen their Christian name with their surname, because they wish to be called by both. The Smiths and the Joneses commenced the practice, but others have followed their lead. The indexer has no means of telling whether in a hyphened name the first name is a real surname or not, and he needs to know much personal and family history before he can decide correctly.

Hyphens are used most recklessly nowadays, and the user has no thought of the trouble he gives to the indexer. If the Christian name is hyphened to the surname, and all the family agree to use the two together as their surname, the indexer must treat the compound name as a true surname. Often a hyphen is used merely to show that the person bearing the names wishes to be known by both, but with no intention of making the Christian name into a surname. Thus a father may not give all his children the same Christian name, but change it for each individual, as one son may be James Somerset-Jones and another George Balfour-Jones. In such a case as this the hyphen is quite out of place, and Jones must still be treated as the only surname. No one has a right to expect his Christian name to be treated as a surname merely by reason of his joining the Christian name to the surname by a hyphen. He must publicly announce his intention of treating his Christian name as a surname, or change it by Act of Parliament. Even when the name is legally changed, there is often room for confusion. The late Mr. Edward Solly, F.R.S., who was very interested in these inquiries, drew my attention to the fact that the family of Hesketh changed their name in 1806 to Bamford by Act of Parliament, and subsequently obtained another Act to change it back to Hesketh. The present form of the family names is Lloyd-Hesketh-Bamford-Hesketh.

With respect to Spanish and Portuguese names it is well to bear in mind that there are several surnames made from Christian names, as, for instance, Fernando is a Christian name and Fernandez is a surname, just as with us Richard is a Christian name and Richards a surname.

9. An adjective is frequently to be preferred to a substantive as a catchword; for instance, when it contains the point of the compound, as Alimentary Canal, English History; also when the compound forms a distinctive name, as Soane Museum.

The object of this rule is often overlooked, and many indexers purposely reject the use of adjectives as headings. One of the most marked instances of an opposite rule may be seen in the index to Hare's Walks in London (1878), where all the alleys, bridges, buildings, churches, courts, houses, streets, etc., are arranged under these headings, and not under the proper name of each. There may be a certain advantage in some of these headings, but few would look for Lisson Grove under Grove, and the climax of absurdity is reached when Chalk Farm is placed under Farm.