One of Johnson's chief aims in embarking on his Dictionary was "to preserve the purity [...] of our English idiom" by including only such words as were used in "the general intercourse of life" or could be found in the works of "those whom we commonly stile polite writers". He derived authority for the inclusion of many words and their meanings from their printed occurrence in books by the "best" writers from "the golden age of our language", and, as testimony, he included quotations from their works (Johnson 1747). These quotations he referred to as his "authorities", but the nature of that authority is ideologically complex and has implications for the treatment of Johnson's Dictionary as an authored text.

Recent developments in literary theory have drawn attention to the notion of authorship. Roland Barthes (1977) proclaimed the death of the author in 1968 and Michel Foucault (1979) has examined the consequences of regarding literary works as authorless, concentrating on an abstract author function rather than a historical, biographical author. As a consequence, there has been a shift in deconstructionist literary criticism from study of authors and authorial functions to study of readership and texts. But other areas of literary study, including scholarly editing, have remained resolutely author-centred, and in the popular mind books are valued because they are written by named authors. The recent interest in historical dictionaries, of which this conference is an expression, has tended to draw attention to these problems of authorship because many of us are dealing with "authored" texts in a genre which would not in the modern world be regarded as literary.

Most people regard modern dictionaries as authorless; they are not really aware of who has written them, and, in fact, for the majority of users dictionaries seem to derive their very authority from their apparent anonymity. Only specialists take an interest in individual contributions to dictionaries, and even then these are not always identifiable. Dictionaries have this status in common with other discourses which do not have an "author" in the generally accepted sense: scientific treatises, educational textbooks, and Hansard (in which the proceedings of the British Parliament are published). What all these texts share is an authority which might be compromised if we knew that they had an "author" who could be named and who had biographically ascertainable character traits. Where authorship obtrudes in this kind of text some measure of objectivity or authority is sacrificed. An example of this kind would be a broadsheet newspaper where the editorial is presented without a by-line and has thereby a greater perceived authority than an identified column by even a well-respected journalist. Quite recently there was a debate about this when the Times Literary Supplement dropped its practice of presenting anonymous reviews and the gist of the objections to this change was that the reviews would become somehow discredited and lacking in authority.

But most texts are not like this; the general direction of development in literary studies has been from the authorless medieval text towards more and more identification of the author as a real person, to the extent that texts come to derive part of their status from being written by a named author. It is this kind of authority that has been called into question by recent moves in literary theory. The case is very different, however, with historical dictionaries: early dictionaries announce their authors on the title-page and their content is often such as to identify the personal tastes and idiosyncrasies of their authors, while more recent dictionaries are presented in a conventional format which minimizes the appearance of personal contribution by the lexicographer.

All this raises interesting questions for Samuel Johnson's Dictionary. The evidence of Johnson's authorship is present not just on the title page, but in entries such as that for "lich" where he adds a comment about Lichfield, his hometown: "Lichfield, the field of the dead, a city in Staffordshire, so named from martyred christians. Salve magna parens." In the absence of much scholarly work on Johnson's Dictionary, the tendency has been for this kind of entry to be highlighted, so that there is a high awareness among the general public of entries in Johnson's Dictionary for "oats", "excise", "whig", "tory", and so on. These are not representative of the Dictionary as a whole, but they do draw attention to its status as an authored text, which is what distinguishes it from modern dictionaries. The problem of authorship is more vexed in Johnson's Dictionary than in other early historical dictionaries because a great deal of the text is apparently not written by Johnson at all, but is constituted by quotations from other texts by other authors. The inclusion of these quotations raises certain questions: to what extent can we say that Johnson "wrote" his Dictionary, and what does it mean to speak of "authorship" of a dictionary? These questions are related to questions about the nature of authority which have been recently raised in literary theory, but does examination of these issues have any relevance for our understanding of a complex text such as Johnson's Dictionary?

The influence of deconstructionist literary theory on literary studies has tended to fragment the unity of the author and to displace the author as the source and centre of the text. The author is no longer God-like and able to dispense "theological" meanings which it is the task of the critic-acolyte to decipher. Correspondingly, the text is no longer considered as the site of final, unified meaning authorized by the author; rather, meaning is located in the complex interplay of reader and text, and the author as biographical subject has been transformed into an intersection of ideologies and discourses. The author may still figure in the text, but only as a kind of fiction, a myth, or an ideological construct.

According to Barthes, the text is irreducibly plural, a complex interlacing of voices and codes which cannot be made to cohere in a single expressive voice belonging to the author:

We know now that a text is not a line of words releasing a single "theological" meaning (the "message" of the Author-God) but a multi-dimensional space in which a variety of writings, none of them original, blend and clash. (146)

There can be no closed, internal, single meaning in the text, he argues, because "a text's unity lies not in its origin but in its destination". The meaning of a text is no longer single and unified by the author's intention, but is volatile and variable with every reading experience. The reading experience itself is governed by what Barthes and Julia Kristeva have called intertextuality. The argument is that every text is an intertext in which other texts are present not just by allusion or quotation, but by the ideological and cultural surroundings out of which they emerge. And the process of reading a text resemanticizes earlier texts so that our reading of them is also transformed.

The image is suggested by the etymology of the word "text": it is something woven, a fabric of citations from past texts and an interlacing of codes and signifiers. Barthes gives as an example reading a text cited by Stendhal:

Proust is what comes to me, not what I summon up; not an 'authority,' simply a circular memory. Which is what the inter-text is: the impossibility of living outside the infinite text þ whether this text be Proust or the daily newspaper or the television screen: the book creates the meaning, the meaning creates life. (36)

Barthes' meaning is characteristically opaque here, but what he seems to be getting at is the notion that intertextuality is more than just the study of sources, influences, and allusions in a text for which there is a recognized method (practised by literary scholars) and a verifiable outcome. The Proustian image is not an authority for him, but a stray, almost whimsical thought, evoked by his reading but not "there" in the text for others to see.

Johnson's Dictionary would seem to be the paradigmatic post-structuralist text. It is not authorial in the sense of embodying the original creative expression of an author revealed in the text, and it presents citations from past texts in an apparently intertextual way. So what justification can there be for wishing to edit such a text, when the governing principle for such activity has normally been authorial intention? What authorial activity is there in the Dictionary which can be recaptured? And if there is an infinite plurality of meaning in a text, why edit it at all? Why not simply present the plurality of published editions? My argument is that the idea of the author which Barthes, Foucault, and deconstructionist critics have tried to dismantle and dismiss is historical: it belongs to a period after Johnson wrote the Dictionary. What needs to be applied to Johnson's Dictionary is a wider notion of authorship than the post-Romantic concept of original genius which is implied in discussions of authorship by deconstructionist critics.

Questions about authorship have tended to be addressed in relation to fictional narratives and all sorts of devices of displaced authorship have been noticed: narrators, points of view, personae, reliable and unreliable commentary, impersonal narration, and so on. These questions have not normally been addressed in relation to works which in the modern world are regarded as non-literary. The writer of a modern dictionary corresponds more or less with Barthes' identification of a particular kind of narrator: an "omniscient, apparently impersonal, consciousness that tells the story from a superior point of view, that of God". The modern lexicographer's assumption of anonymous authority could be seen as God-like, and, by contrast, Johnson's displacement of his role as author into other voices in the text could be seen as comparable to the post-modernist novelist presenting the "story" from multiple points of view. But is this analogy valid and does it inform our notion of the authorship and authority of Johnson's Dictionary?

Johnson had particular intentions in selecting the material for the illustrative quotations, intentions that were partly linguistic (they provided instances of usage), but mainly ideological, having a great deal to do with the extra-lexical purposes that Johnson had in mind for the Dictionary. In the first place, he wished his sources to convey particular ethical views so that he would quote neither from Thomas Hobbes, whose moral principles he thought dangerous, nor from Samuel Clarke, whose style he admired but who held unorthodox views about the Trinity. Vast numbers of quotations are included from High Church Anglican divines who collectively present an orthodox, conservative Protestant theology.

But Johnson also saw the illustrative material as providing an encyclopedia of knowledge, and many of the quotations are included not so much for their suitability as exemplifications of linguistic usage as for the purpose of education and instruction; cumulatively they transmit the current state of learning on many topics. Where Johnson cannot find a quotation that presents the latest state of knowledge, he updates it himself in commentary, as in this example under the word electricity where he quotes from Sir John Quincy:

A property in some bodies, whereby, when rubbed so as to grow warm they draw little bits of paper, or such like substances to them.

but adds this comment:

Such was the account given a few years ago of electricity; but the industry of the present age, first excited by the experiments of Gray, has discovered in electricity a multitude of philosophical wonders. Bodies electrified by a sphere of glass, turned nimbly round, not only emit flame, but may be fitted with such a quantity of the electrical vapour, as, if discharged at once upon a human body, would endanger life. The force of this vapour has hitherto appeared instantaneous, persons at both ends of a long chain seeming to be struck at once. The philosophers are now endeavouring to intercept the strokes of lightning.

Many of the quotations are trimmed and edited in ways which make them more suitable for Johnson's intended purpose. An example of the kind of thing Johnson does with the quotations can be found in the entry for 'law'. Some quotations appear in full form, such as the first from Hooker:

That which doth assigne unto each thinge the kind, that which doth moderate the force and power, that which doth appoint the forme and measure of working, the same we terme a Lawe.[1]

Others are trimmed to make their sense more pithy and to eliminate unnecessary and distracting verbal baggage, as in the second example from Hooker:

Laws politique, ordeined for externall order and regiment amongst men, are never framed as they should be, unlesse presuming the will of man to be inwardly obstinate, rebellious, and averse from all obedience unto the sacred lawes of his nature [...] (Hooker: I, ii, 49; I, x, 70)

Sometimes this process is taken too far and the edited quotation printed in the Dictionary is so far removed from its original and so heavily trimmed as no longer to make any sense. An example of this is the quotation from Thomas Baker's Reflections upon Learning:

Some of the Constitutions have been altered without Judgment, and others in such a manner as betray no little Ignorance on the Compiler: In some the Words are struck out, that determine the Sense of the Law; and again Words added, that give it a new one: One Law is split into Two, and sometimes Two are run into One: the Time and Date are often mistaken, and sometimes the Person. (Baker: xiii, 185)

The irony of this particular quotation is that Baker is objecting to the butchering of texts by editors who change the sense of a text by their alterations.

But there is also an interesting effect arising from the arrangement of the quotations within a single entry of the Dictionary which does not so much alter the sense of a quotation as give it an extra resonance. An example of this is the entry for the word mispend (sic). The quotations are arranged in such a way that they give different inflections to the meaning of the word and tease out some of the religious overtones that the word has for Johnson. He was acutely aware of the Christian moral obligation to make the most of one's talents and to use one's time usefully and this is reflected not just in the selection of quotations for this word, but in their arrangement:

To MISPEND. v.a.
1. To spend ill; to waste; to consume to no purpose; to throw away.

What a deal of cold business doth a man mispend the better part

of life in? In scattering compliments, tendering visits, gathering and
venting news.      Benj. Johnson's Discovery.

First guilty conscience does the mirrour bring,
Then sharp remorse shoots out her angry sting;
And anxious thoughts, within themselves at strife,
Upbraid the long mispent, luxurious life.         Dryden.

I this writer's want of sense arraign,

Treat all his empty pages with disdain,
And think a grave reply mispent and vain.   Blackmore.

He who has lived with the greatest care will find, upon a review

of his time, that he has something to redeem; but he who has mispent
much has still a greater concern.                      Rogers.

Wise men retrieve, as far as they are able, every mispent or

unprofitable hour which has slipped from them. Rogers.

The first quotation, from Ben Jonson, begins with a general complaint about the loss of time in social courtesies, but the following quotation from Dryden immediately associates waste of time with moral guilt, and the survey of past actions has the feel of a death-bed repentance. The quotation from Blackmore deals with the waste of time involved in taking a bad writer seriously enough to bother to reply to his errors, but the position of this quotation in the entry, following the religious associations already evoked by the word, means that the waste of time involved begins to seem morally reprehensible. The two quotations from Rogers have the same effect: the second one appears to be a piece of social wisdom, almost a proverb, but placed after the first quotation, which deals with death-bed reflections, it again takes on a more serious religious and moral import.

This is exactly the effect described by Barthes and Kristeva concerning the intertextuality of texts. Here not only are the source texts altered by being trimmed or edited in some way by Johnson, but they are resemanticized by being taken out of their original context and given another which is entirely different. Not only that, but we would seem to have here the very model of a post-structuralist text. The text of Johnson's Dictionary appears to be fragmentary and disjunctive; it does not have a univocal meaning; the only continuity is arbitrary alphabetical progression: here would seem to be the definitive model of a text which is a "multi-dimensional space in which a variety of writings, none of them original, blend and clash". Yet there is authorial presence in this text. Johnson has organized the "variety of writings" quite deliberately in such a way as to create out of their combination a meaning which is not present singly in any one of the quotations. In addition, he has imposed ideological criteria on the selection of the illustrative quotations such that collectively they convey a particular ethical and theological viewpoint. The notion of the author as original, creative genius obviously does not apply to this text, but if we shake off the anachronistic shackles of post-Romantic conceptions of authorship we can find an equivalent of Johnson's "authorial" activity in the medieval redactor. The selection, arrangement and editing of others' texts can be seen as just as much "authorial" as the writing of a novel or a poem, and in this case it is quite possible, contrary to Barthes' assertion, to see the writer as subject and the author's book as predicate.