DGE: The DGE and Lexicographical Science

F.R. Adrados
Lexicographica 2, 1986, pp. 8-22

The Greek-Spanish Dictionary and Lexicographical Science

Greek-Spanish Dictionary (Diccionario Griego-Español), the first volume[1] of which appeared a few years ago, the second being on the point of publication, is a collective enterprise that attempts to further the lexicographic tradition of Ancient Greek. The characteristics of this project are fully explained in the lengthy prologue to Vol. I, and in the book Introducción a la Lexicografía griega,[2] as are likewise the methods and aims of same. Some of the authors of this work have also published further information to this respect.[3]

However, theory and authors' intentions are one thing and the actual result is quite another in a work which operates with hundreds of thousands of data, whether these come down from a long tradition (with all the problems implied therein) or whether they are new ones and particularly if the work is carried out by an extensive panel of constantly renewed collaborators, Practice is a hard master and it may occur that we did not totally achieve some of the aims we set ourselves: that for example, errors or inconsistencies may have crept in or that the treatment given to certain articles or sectors of the lexicon out to be less complete than would have been desirable or, also, that semantically based theories for the organization or the articles, whatever the circumstances, may have given less perfect results than was hoped for. Yet it could also happen that, on the contrary, precisely this practice of collecting and studying data and of writing up the articles, may have gone beyond the original theory. Moreover, it may turn out that we have made contributions to Lexicographic Science in general, that are worth underlining.

believe the moment has come to comment on these possible differences between the aims and results of our DGE (as it will be called henceforth) so that they will be available to lexicographers in general. These aims and results will doubtless be of interest to them for they concern problems, most of which crop up in any sector of this Science. In fact, we now have a total of fifteen reviews of our DGE.[4] We are enabled by these pages to either thank those who praise us or offer us criticism and contributions of new material or points of view as given in these reviews (and in a series of letters from distinguished lexicographers and diverse specialists received by us); or else to have the chance to explain and defend our procedure, and even to go further into the DGE's contributions which, to our mind, are important and which seem to have escaped the attention of the critics and reviewers. It seems particularly strange to us that there is an almost total absence of references to our procedure in the inner organization of the articles: a subject stressed both in the prologue to the DGE (p.XXXVff.), as likewise in the above-mentioned book Introducción a la Lexicografía griega, pp.228-280). This would seem to be a symptom of the divorce between classical philologists and modern Semantics, which is not a good thing and which we attempt to bridge as far as possible in the following pages.[5] On the other hand, our Vol. II is now at an advanced stage of printing and we should like to anticipate here its contributions to the technique of organizing articles by means of a staggered branching of meanings. We believe this to be just as essential in modern lexicography as are techniques concerning the collection of new data or the interpretation of both new and old data; yet, as I said, this has hardly been noticed.

As I stated at the outset, the DGE does not aspire to open up an independent lexicographic line, although it offers new contributions within the ambit of the lexicographic and semantic Science of our day; it follows the lexicographic lines that come from Stephanus's Greek Thesaurus (which in turn depends on a tradition that goes back to the Hellenistic age) and is handed down through a series of works. Its most immediate predecessor is Liddell-Scott-Jones's great Greek-English Dictionary in its last edition of 1940, reprinted several times and quoted henceforth as LSJ. This is to be completed with Barber's «Supplement» of 1968. Although comparisons are always odious, for it would seem from them that a work that has been and still is important is being underestimated, there is no way of avoiding them when we[6] attempt to make our own contributions or novelties known or when reviewers in turn try to quantify or qualify these contributions.[7] I believe that any Dictionary that is more extensive and more modern than LSJ, whether it be our own or anybody else's that might have appeared, must forcibly surpass it in many respects, which does not mean an underrating of it or denial that it may still be necessary to consult it occasionally, as is also the case with the old Thesaurus and the Dictionary begun by Crönert. Thus, statements such as West's that our and the Dictionary begun by Crönert. Thus, statements such as West's that our DGE «will serve as a complement to LSJ, not as a definitive substitute» could be admitted (what, indeed, is definitive in Lexicography?), and the same goes for Pugliese's wish that the DGE «si affianchi, completo, al benemerito LSJ nello sacaffale dei libri a cui l'ellenista non concede riposo», which flatters us. Really, with all the criticisms of our work contained in the above-mentioned reviews (and obviously, only some of those criticisms that could have been mad have been made), the idea that the DGE in general terms (it could never be said in all points in the case of a work like this) represents some progress within the traditional line we mentioned before, is present in all of them. For us, there are flattering statements such as Dillon's that.

«All in all, it looks as if by the year 2000 we will all be using a basic research tool from Spain, which will be very much against the run of the play hitherto.»

This

, even when our reviewers are not unaware of the risk implied in such an extensive enterprise, undertaken in none too favourable economic and social circumstances.[8]

What does seem clear in that after LSJ ant its «Supplement», there was an imperious need fro a new and fuller Dictionary of Ancient Greek in a modern language. Lacunae such as Christian Greek, proper names (only selectively collected in the DGE), Mycenaean (of which we compile its own Dictionary connected with the DGE by cross-references), the immense amount of new data from papyri and inscriptions, all needed to be remedied. On the other hand, texts already known can also supply new data, particularly when new interpretations yield new meanings of words. In this, the lexicographer either uses commentaries and indices, or he must do the necessary interpretations himself.[9] Our knowledge of Greek dialects is by now also more extensive and this fact also should be reflected. And, in order not to stress each and every point, I would recall that the organization of articles, not only in LSJ but in the whole tradition, should be in accordance with more up-to-date semantic and lexicographic doctrines. Thus, although this is reflected in greater extension of the volume (the DGE is more or less twice the size of the LSJ), it should be stressed that not all is purely quantitative.[10] Lexicography constantly weaves and unweaves its Penelope's loom: we have aspired to not only an improvement in the amount of data by combing through new texts and re-studying the old ones, but also to accuracy with regard to original texts (as far as possible) and to interpretation. All these are traditional techniques which the lexicographer should not disregard. Alongside these, I repeat, is the application of the results of modern Linguistics.

That the need for this enterprise was obvious is to be seen from the two projects underway in Oxford: that for the collection of data for a new supplement to LSJ under the supervision of Peter Glare (with whom we are in contact) and the project for a new lexicon of proper names under the supervision of P. M. Fraser. This latter will surpass our DGE, which in general only collects proper names from literary texts. As regards Glare's work, ours is different in its ambition to incorporate in the Dictionary not only new data but also new interpretations, instead of leaving these for a Supplement. The Supplement formula is practical in some senses, but it has drawbacks that I do not wish to discuss here. In any case, these are two projects which do not interfere with each other (as M. Peters p.86 thinks); that of compiling a new Dictionary, one more extensive than the LSJ (whether compiled by us or by whoever) was obviously needed.

On the other hand, we have never aspired to compiling a Thesaurus in the manner of the Latin scholarship; the obstacles that prevented this enterprise in Germany are already well-known. The Greek lexicon is unbelievably extensive, inexhaustible, we would say, and traditional working and printing methods (which we, in general, follow) are incapable of reducing it to a Thesaurus of the traditional type. The cost in personnel, time and money prevents this.

Yet

, neither do we believe that our work has been made unnecessary by the already highly advanced enterprise of creating a data-bank for the whole of Ancient Greek, the Thesaurus Linguae Graecae by the team directed by Prof. Theodore F. Brunner in the University of California (Irvine). We are in close contact with this project and both our initial List of authors[11] and the Canon of Greek Authors and Works from Homer to A.D. 200 compiled by L. Berkowitz for the TLG[12] were written with the aid of reciprocal consultations. However, the TLG, even though it may some day make all the literary texts machine-readable, even the inscriptions (which the University of North Carolina have taken charge of) and papyri (so far untouched), still has many thorny problems ahead of it in lemmatizing and creating programmes, apart from those concerning the availability of personnel to attend to consultations. But above all, a data-bank is not a Dictionary. The creation of this latter on the basis of the former is theoretically possible (whenever there is a sufficiently well-trained and co-ordinated staff), but we do not know in this case how long that would take in practice. In any case, even if we could avail ourselves of this total Dictionary of Ancient Greek in some uncertain future, whether printed or not, this unsurpassable Dictionary would not eliminate the need fro one of, should we say, medium size like the DGE (after all it is smaller in size than Stephanus's Thesaurus and than the published part of Crönert's Dictionary, although it surpasses them in other respects). (after all it is smaller in size than Stephanus's Thesaurus and than the published part of Crönert's Dictionary, although it surpasses them in other respects).

II

Reading through the reviews of the DGE, plus certain letters from lexicographer colleagues, our own revisions and the study of the new material we are constantly collecting, have all led us to write a Supplement to vol. I (from α to ἀλλά), which will come out as an introduction to vol. II (this latter getting as far as ἀποκοινώνητος). Likewise, both for Lists I, II and III (of authors and words, of inscriptions and ostraka, and of papyri, respectively) and for the Dictionary itself, we have introduced corrections to several types of errors and have added new data. In the case of the Lists, corrections and additions were needed because new texts have been published (above all, new inscriptions and papyri) and also because we have at times thought it highly advisable to substitute the editions used in vol. I for later ones. It should be understood that by no means all of what certain reviewers (above all West) consider to be errors, are considered by us to be such - see below. As regards the introduction to new editions, it is understandable that the change can only be made when it means a quite substantial improvement in the text: a change of edition poses many problems and the Dictionary has to have a certain coherence. On the other hand, new editions that came out when they could no longer be used in vol. II have been left to be included in the new Supplement that will accompany vol. III, in which they will be used.

After making these points, it would be advisable with regard to the reviews in question - although the subject should be extended to debate a series of important questions of principles - to study a series of points that are important for nay lexicographic enterprise comparable to ours. I shall discuss these one by one.

1. The List of Authors:

The List of literary Authors which has basically been transmitted through manuscripts and whose work are referred to in the DGE (2.488, apart from those added in the Supplement that accompanies vol. II), displays certain characteristics that respond to general problems posed by this type of work.

To begin with, it seems clear that in the case of an ancient language like Greek that is also known to us in its medieval and modern form, certain chronological limits should be set when writing a Dictionary. We have established, somewhat conventionally, the year 600 approximately. Of course, we are not unaware of the fact that the evolution of the Greek lexicon form Homer to this date was considerable and that, therefore, the DGE refers to what is called a diasystem and not to a closed system. Neither is there any doubt that there is remarkable continuity with medieval Greek in its form as a literary language or καθαρέουσα. We respect this continuity above all by taking into account Byzantine authors who worked within the ancient philological tradition (the lexicographers, Eustathius, Photius, etc.). In any case, it appears that there was a break in the ancient tradition around the date in question. Wilson (p.211) considers it more logical to also include writers of the reign of Heraclius (610-641), which better marks the end of an epoque: perhaps he is right (and indeed we are following up in practice his suggestion that the Greek papyri of Arabic Egypt should also be considered).

Some lacuna in the list in question have been covered in the above-mentioned Supplement. Another objection which might be made is that the list is «inflated» (thus West p.256) by including authors without literal fragments. This is in fact not true: these authors contain proper names which we have collected and there is a literal fragment from the «Telegonia» (Allen 5.143).

An important point which should be mentioned with regard to the list of authors, and generally, to the names assigned to these latter, is that a Dictionary is not the proper place to discuss problems of authenticity. The names of the authors are those which are given to them in the editions we use, or rather, in the titles of said editions. We do not distinguish between Hippocrates and the Hippocratic writers, Pythagoras and Pythagorians; the «Constitution of Athens» is placed among Xenophon's works because it is edited with Xenophon, although everybody is convinced that it is not by Xenophon. And we are not going to go into whether such and such a speech attributed to Demosthenes or such and such an idyll attributed to Theocritus is really by them or not; we do not even include the author's name in square brackets in cases in which a work is considered spurious. Only in the most obvious cases, for example, do we distinguish Phoc. (Phocylides) from Ps.Phoc.

There would be no need to repeat this - for it is justified in the prologue to the DGE - were it not that it has not been properly understood. The fact that ἄγκιστρον cannot be attributed to Gregory of Nazianzus, an Atticist writer, as Renehan (p.17) says, may be true; but for us, Gr.Naz. is a conventional label that covers all works edited as being by this author. Renehan himself doubts the authorship of Zonaras for a late lexicon: we say the same. The names of authors and collections are a convention for us that simply helps to locate the passages: it is the reader, should the case arise, must later give his opinion on true authorship. How can the author of a Dictionary get mixed up in solving problems of authenticity) He would have to neglect his own work to write a whole library, the results of which would not, on the other hand, be universally accepted. I believe that our solution to this problem is the right one, not only in the case of Greek, but in general terms.

2. Choice of the editions followed

A problem with which the author of a Dictionary is automatically faced when this latter is on ancient texts or problematically transmitted ones, is the following: which edition to follow. According to whether one follows one or another edition, new lemmata are created and others disappear; and within one and the same lemma, one or another morphological from appears. On the other hand, editions of two authors in which the same word appears, offer contradictory variants (see below on ἀγκυλοχήλης / ἀγκυλοχείλης). Moreover, it should not be concealed that the editions offer forms that are mere conjectures (cf. below on ἀβρωσία) and, on the other hand, other forms are hidden in their critical apparatus which may in the opinion of other authors (or each work) following one single edition, this being always the same one, as we do, entails the risk of impoverishing the ancient Greek lexicon.

Nevertheless, we believe that is exactly the right solution. The author of a Dictionary cannot make an edition of each author and work for himself. He is unable to give all the variants in the diverse editions and diverse manuscripts, plus the different conjectures. Nor can he choose according to his own criteria in each case the text of such and such a manuscript or edition. Apart from the fact that this is equivalent to making one's own edition (for which he has neither time nor opportunity), it leaves the reader in the dark as to which text was followed. It has been done like this only too often. A limitation of the data offered is preferable, albeit on the basis of fixed, certain data. Our DGE does not refer to Sophocles: is refers to Dain's edition of Sophocles.

However, this somewhat rigid approach may be softened a little. Several reviewers criticize us for having followed Edmonds's edition of the fragments of the comic writers; it is certainly somewhat fanciful and full of conjectures. If we have chosen that edition, it is because it was in its time the most complete (we would now follow Austin-Kassel's, two volumes of which have just come out). Yet in this case and in others, when required, we give the indication that such and such a form is a conjecture (cj.) in brackets, and we give elsewhere the reading of the manuscript (codd.). The same goes for the case of the inscriptions and papyri: we follow the text of the collection indicated by the respective initial list, but add new readings or restitutions in brackets. We cannot always do this, although it is the available resource for truly needy cases. When there are variants with regard to the lemma in different editions or in inscriptions and papyri, said variants may be included at the beginning of the entry (lemmatic part), in which the form of each of the passages quoted within it is stated (a slip is exceptional such as the one pointed out by Renehan, p.15).

Therefore, the objection that such ans such a form or syntactic construction is not guaranteed for ancient Greek, for it may depend on just one editor, is not at all valid. The careful reader knows, by checking with the initial Lists, which edition is being used; we do not state that this form or construction is true ancient Greek, we merely state that it figures in such and such an edition. From here on, anyone can investigate the facts. We avoid distorting tradition by our own correction or one by an editor whose criteria we could generalize. We merely give the data so that that can be studied.

Another thing is the rightness or wrongness in the choice of the editions of diverse authors and works. Of course, it is hard to please everybody in this field. I have already explained our reasons for following Edmonds for comic writers. I would add that practical reasons have led us to follows the same edition for the whole work of a prolific writer, although we sometimes break this rule if a much improved edition is on the point of coming out (e.g. for Lucianus, Strabo, etc.) Neither can one be changing editions in each volume, although we do introduce certain changes in vol. II, whilst others are planned for vol. III. To say, like West (p.257) that «the editions are not always well chosen» is to say very little. Others like Di Gregorio (p.279) and Irigoin (p.149) are more generous: «dans l'ensemble le choix fait par les responsables du DGE est satisfaisant» are the latter's words. It should be pointed out that the absence of the edition of Aeschylus' scholia carried out by Smith in 1976, which this author mentions, is due to the fact that it did not arrive in time to be used. This is a typical case of the fact that a work of this nature demands such a long period of writing, revision and printing that a long delay must be calculated into the planning until an edition (or collection of papyri or inscriptions) can be used. We only use Smith from vol. III.

One should make up his mind, on the other hand, to the fact that sometimes there is little choice. For example, Peters, p.86, does not like Latte's Hesychius, responsible of the dropping out of the lemma ἀκροτερῆσαι. But there is not a better one.

Sometimes, the reviewers realize that the defects they find are due to the problem of editions. Thus West (p.257), when he points out the existence of the two lemmata mentioned above ἀγκυλοχείλης and ἀγκυλοχήλης, as due to the preferences of the diverse editions. Yet the only thing that can be criticized in this case, as opposed to many others, is that we have not unified both lemmas in one. In this case we indicate at the beginning of the entry that there are two variants and we give the authors and works in which one appears (according to the edition followed).

The case of αἰθερολάμπω, also according to West, is different. This occurs in Koechly's edition of Manetho, which we follow. According to West, LSJ is right when it gives a different lemma αἰθερολαμπής, one which must have come from either a conjecture or an older edition. In West's opinion, «those lexicographers did not follow editions so blindly; they may have consulted Koechly», which is not true, as this lemma comes from the Thesaurus, «but had the sense to dismiss the supposed verb as an impossible creature». To my mind, if the edition says αἰθερολάμπει which can only come from αἰθερολάμπω, this lemma should be given; only if it were such an «impossible creature» as West says (and I personally do not believe this), some information should have been given in brackets (such as «perhaps corrupt» or «perhaps for αἰθερολαμπής»), as we often do, above all in lexicographers' lemmas. However, it seems incorrect to me to invent a lemma without justifying where it comes from.

I repeat that it is not the lexicographer's task, during the writing phase of a Dictionary intended to collect and classify material, to introduce forms which are only found in manuscripts. Renehan, for example, says (p.16) that ἀβελτηρία with eta is not only in Chrys., but also appears in manuscripts of Plato; but it is not our task to give a complete inventory of the passages in which each lemma appears nor even less to note the appearance in manuscripts of forms which in the case of other authors are accepted by their editors. Neither is it our task to stipulate that a lemma such as ἀβρωσία, which figures in our edition of E. Hippol., comes from a conjecture (as opposed to Wilson, p.212). We are not concerned either with the restitution of, for example, the original form of ἀηγύβερσον Zonar., or with indicating that, as somebody thinks, the examples of ἀδάματος are editor's errors. Exception may perhaps be made such as introducing lemmata based only on some variant from a manuscript and pointing this out (thus ἀθαλασσία in our Supplement). But see below for further details on this.

What is more serious is that certain reviewers do not realize that the presence or absence of certain lemmata in the DGE depends on this type of problem. Thus, Georgountzos (p.397) criticizes the fact that ἀάρης is missing from the DGE and quotes Et.Gud. This is because De Stefani's edition, which is the one we follow, gives ἀάνης, which is our lemma. He also criticizes the fact that we do not include ἀκαθόπλιστος; but this is just a v.l. of Sch.Il. The same could be said about ἀκρωτερῆσαι and αἰσχόω (in Peters, p.86). We could continue in the same vein.

3. Lemmata and lemmatization

It is in any case quite clear that the Greek Lexicon is practically inexhaustible and that new lemmata can always be found, either in our reviewers, in the new material which is published, or in renewed study of older material. I have already said that in the Supplement to vol. II numerous new lemmas are included.

However this may be, it should be pointed out that the inclusion of a new lemma demands careful previous study. Against what Georgountzos says, ἄαται is not missing because it is in ἀάω: nor is ἅλαι because it is in ἅλη; nor is ἄα missing because it is a false nominative invented on the gen. ἄας which we do include. As to ἅβα, the most one can say is that it should have been introduced as a reference to ἥβη. Ἀθαναϊσταί is included, albeit with the Ionian spelling Ἀθηναϊσταί; some indication should have been given of the Doric form (of Rhodes).

Another problem, which is the opposite of this latter, is whether and how forms from later authors can be introduced, above all forms from lexicographers, which are possibly corrupt. We are in general criticized for having used too many of these forms. I still maintain my stance that they should be introduced; what today may seem erroneous or corrupt may one day be interpreted as genuine. We cannot hide material which could eventually become useful.

Thus, I cannot help differing from such a distinguished lexicographer as Renehan with regard to a series of lemmata which come from a lexicon supposedly by Zonaras. Renehan gives very good reasons for thinking (p.181 ss.) that ἀαχῶρα, ἀγγυιΐσας, ἀγλαόδημος, ἄθη, αἴγαρος and ἀκτάλη are corruptions; respectively for ἄχωρ, ἀγκυρίσας, ἀγλαότιμος, ἀθήλυντος, αἴγαγρος and ἀκταίη. To my mind, if we wish to be coherent and show the reader the existing evidence, we should maintain the lemmata of our edition. Of course, some information in brackets would have been desirable (as in so many other cases) as to a plausible interpretation. I would say the same for ἄγουρος, ἀγορία, ἄγορος, which Brixhe (p.2) rightly points out as forms under which ἄορος is hidden (with a trace of the digamma). Or else for ἀθρηνᾶ and ἀκρύβδηλος, the fact that these are words invented in order to justify etymologies (as West points out on p.257) should have been mentioned at most in brackets (as we have done in other instances).

For the rest, I have no doubt that at times efforts at interpretation of anomalous forms by lexicographers could have been taken further: cf. a good example in Georgountzos, p. 396 ἀαναίμα.

I therefor support including lexicographers' forms just as these latter give them, except when they display obvious alterations (of iotacism, etc.) or when they can be interpreted as declined or conjugated forms of known lemmata. We even include lemmata from Hsch. and other lexicographers preceded by a crux in the respective editions: I believe this is correct procedure. Moreover, I am bound to reply to certain critiques and repeat what is already stated in the prologue to the DGE in the sense that critical symbols used in Latte's edition of Hsch (asterisks, square brackets) have been suppressed. They are ambiguous because Latte uses them in a different sense to the habitual one. And whether the lemma in question comes or not from Cyrillus or is in such and such a group of manuscripts, etc., is the kind of information which the reader can glean from Latte himself, for he is referred to Latte by the DGE.

In other instances, the problems of lemmatization are concerned with the criteria according to which dialectal forms have been hierarchized, with whether or not diverse verbal or adjectival stems are unified in a lemma, with whether a habitually middle verb that at a later date appears as an active form is given as middle or active, etc. Only few remarks have been passed on this by our reviewers whilst it is nevertheless an important and difficult subject. At any case, I should like to criticize Peters' opinion that if we include in one and the same lemma ἀγάομαι and ἀγαίομαι, ἀγοραῖος and ἀγορήϊος, ἀλέομαι and ἀλεύω, that is due to our supposed lack of understanding of phonetical problems: they are not «lautliche Varianten». I think the opposite is exactly true. He may be right, on the other hand, on ἀλήθεια and ἀλάθεα.

Finally, I would draw attention to another important subject, one mentioned by West (p.256) only too lightly. For him, it is a «curious policy» to give from example a lemma ἄγαγας from Petronius' agaga and in general to introduce material that comes from Latin authors. Yet, although there is obviously a problem of limits (as to what extent a loan from Greek to Latin preserves the old meaning or has developed a new one), it is quite clear that the Latin authors give Greek words that are not attested in Greek (as the one mentioned above); and above all, they record Greek words from a date prior to our Greek documentation (in the case of Plautus, Lucilius, etc.). As I have already said, they also give meanings which are probably Greek. On the contrary, I believe that one important contribution of the DGE has been to introduce a large number of Latin borrowings to Greek, above all in administrative and military vocabulary.

In any case, if we have overdone things somewhat in this case, it is once more through a desire to make as much material as possible available to the reader.

4. Morphology and Etymology

The

boundaries between a Lexicon and a morphological (and syntactic) treatise are shaky, for several reasons that I am not going to go into here.[13] We have taken, shall we say, a practical approach. We do not give the morphological forms we would call «regular» but only the anomalous ones (archaic or dialectal outside regular, late or deviant conjugation or declension). Here, I repeat, there is no theoretical justification but the need to set certain limits: to help the reader without overburdening the work with well-know data. As this was already stated in the prologue to the DGE, West's and Peters' remarks (pp. 257 and 87 respectively) on the lack of regular forms of αἴρω or αἰσθάνομαι does not seem fitting. On the other hand, remarks on the lack of certain forms (declension of Ἀαρών, Renehan p.15; comparative of ἀκροδίκαιος, Brixhe, p. 2, certain epigraphical forms, Peters, p.87) are indeed relevant. I have no doubt that others could be pointed out.

A few words, finally, about etymology. I think it is my right to present my ideas and to try not to let them to remain on this side of the Pyrenees, as Peters (who gives a very prejudiced view of same) suggests. I hope he, also, would prefer not to let his own «Laryngaltheorie» to remain by the Danube.

5. Documentation, translations

Obviously, the documentation on our lemmata could be almost indefinitely enlarged: in the Supplement to vol. II, we add some material to that of vol. I when it is particularly significant: quotations that are older of those that belong to different genres and epochs and, of course, quotations with new meanings. It is, of course, possible to systematically increase the documentation of this and successive Supplements. But it is clear that Hellenists will thank us for collecting in our Supplement the new documentation given for a series of words, above all in Renehan's review (p.10ff.), as well as his contribution of older material than ours on certain words (ἀθανασία Renehan p.13). In any case, we have carried out our task selectively and nobody could miss a record from the «Titanomachia» for ἄγω (cf. Führer, p. 195) when there are similar ones from Homer. (cf. Führer, p. 195) when there are similar ones from Homer.

Naturally, one should also be grateful for the discussion of certain translations or for the suggestion of new meanings: cf. on ἁλιτενής in Wilson (p.212), ἀγών in Longin, in Renehan (p.14), ἄγω in Hes. in Führer (p.195), in Hes. in Führer (p.195), ἀγείρω in Plato in Irigoin (p.150), ἀγορά in Pugliese (p.78). All of this will be included in our Supplements. Naturally, things are in other instances somewhat more obscure to my mind; thus, in the case of ᾄδω in Strab. according to Wilson (p.212), etc.

On the other hand, I should not like to leave out the fact that the establishment of meanings is a delicate matter, and, to a greater extent, a subjective one: if ἀβελτηρία in Chrys. deserves a new number within this lemma (2 depravación) in the DGE, or if one sole item would be sufficient as Renehan believes (p.16) is a moot point. But I shall return to this matter of the organization of entries which has been somewhat neglected by the reviewers.

These are the points in which I believe re-examination of certain criticisms could be of general interest and become enlightening. Many other points on lemmatization, on diverse borderline problems, formalization, etc., would deserve some discussion, perhaps; but among the lacunae which we have found in the reviews there remains that which concerns the organization of entries and their semantics, which should, to my mind, be preferably enlarged upon. I shall do this below in an attempt to expound our contributions in these fields: as I have said, not only on the strength of vol. I, but also of vol. II, forthcoming.

III

Here, I am not going to justify the semantic and lexicographic principles of the entries to the DGE: they may be seen in detail, as I stated at the beginning, in several chapters of mine in the Introducción a la Lexicografía griega, as well as in diverse above-mentioned articles and in the Introduction to the DGE itself. What I do whish to point out here, as I also announced, is to explicate this doctrine through a few examples of its practical application, for general formulations are one thing and seeing how they are adapted (or how they have to be modified) to a not always docile material is another.

On the other hand, the doctrine is in general terms common property today in our Science. The DGE is a bilingual dictionary and the guiding principle for writing up its entries should therefore be to organize meanings from the point of view of the target language, in this case Spanish: for example, it is a question of deciding in which circumstances (which must be detailed) the verb ἄγω can be translated as `llevar' (with movement away from the subject) and in which others as `traer' (with movement towards the subject), and the same goes, to set a second example, for deciding upon the circumstances in which εἶμι is to be translated as `ir' or as `venir'.

I believe that the principle itself, by which the target language equivalent(s) should help organizing the entry in general and the sequence of the senses of the entry word in particular, is irrefutable; however, it is not always equally easy, or possible, to follow it. If there are formal elements in the contexts by which one can support a certain organization of the entry, the task is easier.

Such contextual, formal elements, can be established, at least theoretically, in three different ways:

a) By the paradigmatic context. Παῖς has a different meaning when opposed to ἀνήρ (`child'), to πατήρ (`son'), to δεσπότης (`slave'), ἀνήρ is in turn different if it is opposed to παῖς (`mature man'), to γυνή (`husband'). These oppositive contexts may appear in the syntagmatic one (or context in the usual sense) or may not appear; in any case, the syntagmatic context can lead us to the same definitions: an oppositive context corresponds to a syntagmatic one. This is to be seen particularly clearly in the prepositions: oppositions such as that of ἐκ and ἀπό only occur in certain contexts; it is more delicate to define the contexts in which εἰς, πρός, παρά etc. with acc. are opposed. On the other hand, the oppositions are of several types as is known (privatives, etc.).

b) Definitions of meanings by the syntagmatic contexts is, then, redundant at times with the paradigmatic definition. Both may be given or only the later, when it is simpler; but normally in our DGE it happens that the explicit definition is the syntagmatic one. It is easier to give and at times there is no paradigmatic opposition at word-level. However, these syntagmatic contexts may, as is known, be of several types: there is the extra-linguistic context in which the word is used, which concerns the type of language and linguistic register (in such or such science or technique, colloquial usage, dialect ...); there is the general context («with movement from» or «towards» the speaker, «with positive» or «negative evaluation», «with temporal reference» or otherwise, «at the ethical», «logical», or «epistemological levels», etc., etc.); there is the syntactic context (with acc. of this or that type, for example); there is that based on sub-classes of words (with verbs of movement, accipiendi, of process ...; with subj. of person, thing, or abstract noun ...); there is the lexical context (sometimes, only a very specific combination of words establishes the meaning). These contexts occasionally appear subordinated to each other, thus creating semantic sub-classifications, but it also happens that only a combination of contexts (e.g. a syntactic one and one of word sub-classes) determines the meaning. There are even more complicated cases: the meanings of the prepositions are established (besides through oppositive means) by a syntactic element (the case they govern) and three word sub-classes (those of subject, verb and government). All these indications are given in plain type when they are given in general terms. But at times, it is preferable to give a Greek word (or the least possible number of same) as a significant context, without going into which sub-class it belongs to (or without exemplifying a sub-class that has previously been indicated in plain type).

c) There is one more type of definition: that based on an analysis of the words. For example ἀνα-, on the basis of its different meanings, creates different meanings in its verbal compounds («with upward movement», «backwards», etc.). Another example is that morphological analysis sometimes has semantic relevance as certain meanings require the plural (thus in the case of ἄνθρωπος, ἀνήρ) or the middle voice, or a certain verb stem. The principle upon which we base our Dictionary (one which is really widely used, both now and previously, but to which we attempt to give coherence) is the following: we do not attempt to give for either a word or its meaning the total sum of semantically significant data (contexts, etc.) but seek a strategy for offering those which are most easily extracted, expounded and appreciated by the reader from the texts. This strategy sometimes provides staggered data which define the successive bifurcations and branches of the semantic tree: A/B/C ..., I/II/III ..., 1/2/3 ..., a/b/c ... However, any of the meanings may also have a negative formal characterization (lack of data which define a different one). At other times, this strategy, at a given point on this semantic scale, can provide complementary paradigmatic, syntagmatic or analytical data; at times more than one of each of these models.

A Dictionary is an eminently practical tool which leads the reader who knows how to use it from a Greek word to a Spanish one (in this case); this with indications of the circumstances in which the former may be translated by the latter (which may also be a group of words). Therefore, there must be a strategy in the choice of the defining features. Yet there must in turn be a strategy in the expounding of the meanings. Thus, in the classifications of elevated hierarchy a term may be given which is later specified through semi-synonyms in lower classifications (`de', `de, desde', `de, a partir de', etc.) or opens upon into a gamut of semi-synonyms within one meaning (each one is followed by its exemplification and the block thus constituted is closed by a semi-colon).

On the other hand, a Dictionary does not attempt to offer that which the language (or our knowledge of same) does not offer. It does not give non-existent lemmas: present forms of verbs which have only an aorist or a perfect, the feminine forms of adjectives fro which we do not know whether they possessed a special form for same or not, indications of genders or accents which are dubious. It explains the truly existing formal variants and makes quite clear which are those that figure in each example. However, it may analogously occur that a classification of elevated hierarchy (A or I, for example) has not a term for translation in Spanish which suits all the examples. It would the suffice to give a paraphrase and the above-mentioned formal indications.

Formalization, that is, the establishing of the conditioning features of the different translations at least at the level of meaning, is naturally more or less easy to carry out according to case. In general terms it is easier in the field of the verb on account of its wealth of constructions; the organization of nouns (and adjectives) related to the verbs is usually done on the model of these latter, but it is not always thus. Sometimes our material is insufficient not merely to carry out formalized semantic classifications but also to translate a word. On the other hand, the establishment of sub-classes of words poses problems. Sometimes, the opposition «of persons»/«of things» and other of the kind are semantically relevant, sometimes not (and the same must be said of the syntactic oppositions). The lexicographer and the semanticist must be on the look out if they are to find in each entry the strategy which best suits their material. They must be also aware that while a word can have a typical syntactic pattern and words with which it typically collocates may form distinct (semantic) sub-classes [human, abstract, etc., as the case may be], it may also occur «absolutely», i.e., without such formal binders, but still with the same sense. (We put contexts of this type at the end of the respective part of the entry). In such a case, it is the broader context in which the entry word occurs that must be used as the decisive element instead of the usual narrower one.

Of course, there are criteria which are repeated and which even establish throughout the whole Dictionary (and particularly the DGE) certain habits or norms. Very frequent are classifications of meanings on the strength of oppositions of the general context such as «transitive»/«Intransitive», «concrete»/«abstract», «possible»/«passive»/«active», etc., etc. But to give the meanings of an entry in this order is conventional and may fail.

It should be stated at this juncture that the principles of organization of the entries in the DGE are neither simply logical (in the common sense meaning of the word), nor exclusively chronological; however, there frequently arises a situation in which various points of view can be applied, neither of them with an absolute force; in such a situation, we frequently proceed by the chronological criterion.

When the semantic classification leaves a margin (when, for example, the A/B or I/II is conventional), we wither follow the tradition of making the passive meaning (in adjectives in -os) come before the active, the concrete before the abstract, the literal meaning before the figurative one, etc., or else we use the chronological criterion (older meanings before newer ones). We also use the latter secondarily within one and the same meaning. But there may be interferences. At times, the active use (in adjectives in -os) is by far the oldest, at times the figurative use (in nouns), etc. When the chronological differences collide with a simple convention, we follow the former criterion and disregard the latter.

Of course, as this is a panchronic dictionary, there is nothing odd in the fact that certain meaning (as likewise certain lemmata), should belong only to one period (and sometimes, one genre or dialect) or Greek. In these cases it is sufficient merely to point this out.

A subsidiary criterion that could be useful is that of transformations: the indication, within the adjectives, of diverse substantive uses; within the prepositions, of groups composed of article + prep. + noun which are explained by transformation on the basis of identical groups with a verb; within the transitive uses of the verbs, of passive transformations which only modify the meaning from this point of view.

In the Introducción a la Lexicografía griega (p.268 ss.) I already gave a few minimum examples of the organization of entries along the lines of these syntactic criteria; that of ἀμύνω was based on a combination of elements: morphology (voice), government of the verb (acc., dat., gen.), word sub-classes (of person and thing); that of ἀκμάζω was based almost exclusively on this latter criterion (fruits, persons, animals, cities and states, abstracts indicating processes or activity, etc.). There are the criteria which are the key to whether ἀλλά should be translated as pero or sino, sin embargo, etc.; to whether ἀμύνω is rechazar, ir en ayuda de, corresponder a alguien con algo, etc.; to whether ἀκμάζω is estar maduro, estar en el momento más violento, estar avanzado, etc. This means definitely discarding the type of dictionary which indiscriminately gives a long series of meanings, or which by isolating meanings in groups accompanied by an exemplification fails to explain the formal bases of the latter. Although such and such an entry may offer classifications close to those given before intuitively by other dictionaries, it is not often thus. I believe that this is the aspect of the DGE which displays most novelty, to the same extent as or even more than the increase in material it offers.

However, in the above-mentioned Introducción exemplification was minimal as regards this. I give here some new examples: as I have already stated, they frequently come from vol. II of the DGE which is as yet unpublished. Not only should exemplification be given as such but some conclusions should be drawn as to the different types of strategy imposed by the different entries or lemmas.

shall begin my exemplification with a few verbs, which are easier to formalize than nouns and adjectives, as I said before. I shall give first a schema of ἀλλάσσω.[14]

I: with acc. only 1 with regard to what had been established cambiar, modificar, alterar. 2 cambiar, modificar, variar shape or colour; variar, alterar voice or sound. 3 of place cambiar de, variar de. 4 according to the deal trocar or pagar.
II: with acc. plus another construction 1 with acc. and gen., also with acc. and ἀντί + gen. intercambiar, cambiar one thing for another. 2 with acc. and dat. of pers. devolver, pagar; fig. midd. voice with gen. valorar.
III: intr. with plur. or dual subj. 1 alternar, turnarse. 2 with gen. diferir. 3 in midd. voice reconciliarse.
IV: intr. uses in midd. voice or rad. aorist. 1 modificarse, alterarse ἀρεταί. 2 variar τὸ χρῶμα. 3 with movement mudarse πρὸς ἕτερα χωρία.

It may be observed that I do not give concrete examples here unless it is advisable to indicate the subject by means of the Greek word. I and II are transitive, III and IV intransitive; the oppositions I/II and III/IV are clearly indicated in the form either by syntactic constructions or by the morphology of the subject or the voice or the stem of the verb. The lower classifications are based on general contexts as well as on syntax; word sub-classes also intervene occasionally (II 2 with dat of perss., cf. I 4, IV 1 and 2).

As can be seen, through a combination of a series of resources, wide semantic dispersion is achieved; the exemplification which «fills in» the above-mentioned schema in the DGE makes this clearer. I should also like to remark that in this case there is asymmetry with the noun ἀλλαγή: a classification is required here «with»/«without economic meaning» submitted in turn to several subdivisions.

This very disconnection between verb and noun, which is on the other hand rare, is given in ἀλληγορέω/ἀλληγορία. To begin with the verb, I give here the schema for its organization in meanings:

I: with subj. of pers. in general. 1 abstr. and with int. acc. decir con doble sentido or intención common in archaic poets and philosophers; among Hellenized Jews when referring to the Old Testament. 2 with two acc. + adverbial phrase dar el sentido alegóricoalegórico τὸν λογισμὸν Ἀθηνᾶς ἐπιφάνειαν, νοῦν ὡς ἵππον. 3 only with external acc. referring to subjacent meaning aludir alegóricamente a.
II: 3 with subj. of word or story tener sentido alegóricoalegórico, ser una alegoría de ἡ δεκάλογος ... τὰ θεῖα μυστήρια; in midd.-pass. voice with gen. or prepositional phrase (ἀ. περὶ + gen.)
III: 1 with subj. of pers., of exegetes entender or interpretar alegóricamentealegóricamente.

What is worth noting about this verb is the fact that its fundamental classification depends on the different types of subject; the other actant of the verb is the main key (together with levels of chronology and genre) to the subdivision of Roman figures into Arabic ones. As may be seen, things are quite different in these two verbs. As far as the noun ἀλληγορία is concerned, I have already stated that its classification is different:

I: in general 1 palabra or pasaje con significado doble or oculto when speaking of political matters, mysteries, veiled threats ... 2 specialised uses in gramm. and rhet.
II: in the exegesis of poetical and religious texts. 1 in the archaic epoque palabra or serie de palabras con significado oculto que es considerado el verdadero, alegoría. 2 in biblical exegesis, the same meaning.

As opposed to a general use, there is here another use defined by the type of texts. It should be noted that there is no semantic difference between II 1 and 2 which could have been unified.

The case of ἀντιλαμβάνω is somewhat different, for the A/B classification is based on the morphology and B I/II/III on word sub-classes.

A: in general in act. voice with acc. recibir a cambio.
B: in general in midd. voice and with gen. I 1 in the physical sense, with gen. coger, agarrar (there are several subordinate meanings, the last one with gen. of pers. is cautivar). 2 with gen. of regions, the sea, booty apoderarse de. 3 with gen. of sensations captar, percibir. 4 abs. of plants agarrar.
II: of abstr., with gen. 1 tomar sobre sísí, hacerse cargo de τοῦ κοινοῦ τῆς σωτηρίας abs. with adv. ponerse en acciónacción. 2 ocuparse de, dedicarse τῆς παιδείας. 3 aspirar τῆς ἐλευθερίας. 4 abs. reclamar. 5 admitir τῆς ἐλαττώσεως.
III: of pers. or collectives 1 in the political context ayudar ἡμῶν also in general. 2 of horses and men contenerse. 3 in negative contexts atacar, criticar ἡμῶν.

Apart from my initial observations, I should like to pass a few more remarks. In the first place, the semantic unity of B I (`coger' in general terms) and B II (`hacerse cargo') is quite clear. Secondly, the negative definition of B I in opposition to II and III. Thirdly, the hierarchization of word-sub-classes which intervene in the I/II/III classification and in the subordinate ones of these latter. Fourthly, the interest of the the general context (cf. B III 2).

I will leave verbs here and go on to the pronoun ἀλλήλους in order to give some idea of the importance of grammatical contexts with others in the organization of certain words.

A: reciprocal uses I as government of verbs 1 with acc. a) without prep. and acc. uno(s) con otro(s) (according to class of verbs and subjects). 2 with dat. proper. with verbs of «giving», «saying», etc. uno(s) a otro(s); b) with sociative dat. uno(s) a otro(s) or con otro(s) μίσγεσθαι; c) with prep + acc. and verb. of movement uno(s) contra otro(s) ἰέναι. 3 with gen. a) with verbs of action uno(s) de otro(s) γευσόμεθα; b) with verbs of separation unos de otros ὁρίζουσι; c) with prep. and gen. uno(s) de otro(s) γηθοσύνας ... δέχοντο.
II: depending on adj. or adv. 1 with dat. uno(s) a otro(s) a) with predicative adj. ἀγνῶτες θεοὶ ἀλλήλοισι πέλονται; b) with adj. of likeness εἰκέλω ἀλλήλοιιν; c) with adj. or adv. of proximity πλησίον. 2 with gen. uno(s) de otro(s) a) with adj. or adv. of proximity or distance ἀντίον; b) with ἄνευ el uno independiente del otro. 3 with πρός and acc. de uno contra oro θρασύς.
III: depending on nouns 1 with gen. 2 with dat. (same meanings).
B: non-reciprocal uses I with verbs of state or in the perfect. 1 with παρά + dat. uno junto a otro, with ἐπί uno sobre otro. 2 with verbs of «putting» uno junto a otro, seguidos.
II: in gen. de unos y otros νεκρούς δ' ἀ. ἔρυον.

Here, as I have said, classification is purely grammatical: reciprocal uses, word-classes of the governing element, case governed; then, the word sub-classes come into play. But I should warn that in a lemma such as this, differences in translation are often minimal between one item and others; on the other hand, there are sometimes differences within the same item according to word sub-classes of the governed element. This could be seen better if we had not summarized the entry so much. In cases such as these the interest of our treatment of words surpasses semantics to give an image of the contexts in which words are used: above all of the syntactic contexts in the present example.

I will now pass on the exemplification of nouns: a noun is usually dominated by the sub-classes of agreeing adjectives or genitives, as likewise by general contexts of the concrete/abstract, proper/figurative type and other more specific ones which refer to the register of the language and specialized languages. Oppositive contexts occasionally help and also transformation on the strength of verbs and even of adjectives and nouns. A relatively simple case is that of ἄνθος.

I: 1 proper use flor. 2 derived use espuma, erupción ...
II: fig. 1 of temporary processes culminación ἥβης. 2 of pers. with gen. partitive lo mejor de Ἀργείων. 3 of abstr. and things in gener. with epexegetic gen. flor, exquisitez, belleza of love, happiness ...
III: specialized uses color χροιᾶς, brillo of metals ...
IV: bot. ἄ. πεδινόν = ἀνθυλλίς.
V: zool. a bird (perhaps verderón).

I do not think that further explanations are needed. Let us now look at ἀλήθεια, in which classification depends to a great extent on general contexts (I am completing certain points put forward in Introducción, p. 274).

I: in the context of verbs of «saying» verdad.
II: referring to actions, pers., things 1 in the field of ethics, as a virtue verdad, veracidad (τιμάν, etc.). 2 in the epistemological field verdad, conocimiento verdadero often opp. δόξα, ζητεῖν, etc.). 3 in the ontological field verdad, realidad (ἰδέμεν etc.). 4 conceived of as an objective reality verdad Ὀλυμπία, δέσποιν' ἀληθείας.
III: hypostasized Verdad.

In such a case, the morphological and syntactic contexts hardly function. On the other hand, the translation is or may always be verdad, although we use the device of giving shades of meaning so that the philosophical interest of the entry is increased.

Frequently, at one position of the entry an oppositive definition is introduced, usually in a redundant form. This means is more important in other instances as was stated before. Thus in ἀνήρ, which I summarize briefly as follows:

I: opp. γυνή 1 hombre, varón. 2 marido.
II: 1 opp. παῖς hombre older than παῖς. 2 fig. hombre, valiente.
III: neutralized uses (for men and women) 1 hombre opp. θεός. 2 hombre as a social being, ciudadano; hombre, habitante.
IV: semi-pronominal uses 1 with indefinite value, with titles or professions and group denominations (ἄνδρες Ἀθηναῖοι, etc.) 2 in the generic sense uno πᾶς δ' ἀνήρ = πᾶς τις. 3 in distributive uses with κατά, etc. cada uno. 4 in deictic uses with ὄδε: ἀνήρ ὄδε (= ἐγώ).

In any case, it should be noted that a term may enter into several oppositions and that one or more of these may be neutralized, moreover, these contexts are finally defined syntagmatically. It is also important to point out that one and the same translation hombre is repeated in several items, specified in each one with other terms. I would finally add the example of ἄνθρωπος:

A: I in general 1 plur. non generic, generically without an article hombres, personas, gente πολλῶν ἀνθρώπων ... ἄστεα). 2 plur. generic., generally without an article los hombres, los seres humanos, la humanidad (ἀδικότατον ... πάντων ἀνθρώπων). 3 sing., non-generic, rarely with an article ser humano, persona (ἄ. ἄνθρωπον γεννᾷ). 4 sing. generic, frequ. with art. el hombre (βραχύς ... ὁ βίος τοῦ ἀνθρώπου).
II: grammaticalized expressions, equivalent to an indefinite pronoun uno, uno cualquiera (ἀ. τῶν υἱῶν Ἰσραήλ).
III: pejorative and in colloquial language in order to avoid ἀνήρ, γυνή or the proper name (several formalized meanings: amigo, individuo, esclavo).
IV: specialized uses plur. varones opp. γυναῖκες (τῶν ἀνθρώπων τοὺς ἀχρειοτάτους); ὁ ἄ. el varónvarón by opp. to ἡ γυνή, ἡ ἄ. la mujer.
B: medic. ὁ ἄ. el cuerpo humano.

The play of sing. and plur., use or absence of an article, either neutral or polarized use, generic or otherwise can be seen; all give a rich ensemble of meaning that for the rest, most often include the term hombre.

I now pass on to adjectives in which a variety of combinations and diverse contexts also appear. Thus, however much ἁγνός may in general terms mean puro, santo, an organization based on word sub-classes of the nouns to which it refers contributes important points for Greek religion:

A: of gods and pers. I 1 sacro, santo frequ. of female agricultural deities and women related to them. 2 of gods sacro, santo.
II: of pers. 1 particularly of men inocente de crimen, libre de sangre. 2 puro, santo, que no rompe la ley religiosa.
III: related to sexual taboos, frequ. with gen. γάμων or similar 1 puro, no contaminado por relación incestuosa. 2 generally of women puro, purificado después de las relaciones sexuales; generally puro, casto.
B: of things I consecrated to the gods 1 in offerings and sacrifices no sangriento, incruento. 2 of that pertaining to the gods sagrado, santo of feasts, holy places, etc.
II: in a wider sense, with sacral remains 1 puro, incontaminado of water, light ... 2 of song or the voice in religious feasts puro, santo.

As may be observed, translations could be given that in other instances are vague or generalizing and have little to say about the Greek sense of what is sacred.

The general and specific contexts entailed by a word such as ἄλογος are quite different:

A: ref. to proportion and number I arith. and geom. irracional.
II: rhythm. and music irracional of intervals.
III: in general 1 no computado, no tenido en cuenta ἡμέραι. 2 que no paga ἐρανισταί.
B: with regard to «giving reasons» or «explaining» I in gen. 1 no fundamentado, absurdo, infundado. 2 inexplicable, no justificable.
II: phil. and science in opp. to a rational principle 1 of explanations, doctrines no científicocientífico (in other contexts ilógico). 2 of pers. and abstr. in relation to man irracional, no regido por un principio racional (παῖδες, θυμός). 3 inexpresable (τὸ μὴ ὄν). 4 of linguistic elements que no da explicación de los objetos nombrados (e.g. a syllable).

I believe that this makes clear the wealth of classificatory possibilities reflected in different translations (although some, such as irracional, occur in several items).

I will end my exemplification of adjectives with two examples, one centered on the opposition criterion (ἀλλότριος) and the other on the morphological one (ἅπας).

I: referring to individual property, frequ. opp. ἴφιος. 1 de otro, ajeno a) of property, wife or children and parts of the body; also n. plur. subst. τὰ ἀ. 2 fig. como si no fuera propio τοῖς σώμασιν ... ἀλλοτροτάτοις. 3 in relation to mode of being ajeno, que no le corrresponde οὔκ ἀ. ἄτην.
II: in relation to wider social circumstances, frequ. opp. οἰκεῖος 1 in relation to the family circle and friends extraño, no pariente, alejado, desconocido γυνή etc. (several opps.: συγγενεῖς σφέτεροι); οἱ ἀ. extranei heredes. 2 in relation to the country, city ... extraño, extranjero πόλις, ἀνδράποδον. 3 hellenistic phil. ajeno τῷ μὲν ἀστείῳ ἀλλότριον οὐδέν; in general with gen. extraño, ajeno a.
III: in contexts with a situation of hostility 1 in general contrario, hostil, enemigo ἀλλότρια φρονεῖν tener sentimientos hostiles, ὁ ἀ. el Enemigo. 2 phil. opuesto, contrario with gen.
IV: of physical and physiological phenomena 1 provocado por un agente extrañoextraño φῶς. 2 of words inadecuado.

It should be noted how groups of meanings I and II are defined both in general context and by different oppositions. Their subdivisions (and III) are supported by general contexts (subordinate to the former in I and II) an specific ones (Phil., Hellenistic Phil., etc.).

I: plur. 1 of pers. or gods todos, todos juntos, sin excepciónexcepción. 2 of things and abstr. todos, toda clase de ἀγλαίαι, ὀδύναι.
II: sing. 1 a) todo, completo, entero for places, cities ...; b) of time entero ἐνιαυτός; c) for collectives and living beings todo entero λαός. 2 for food, drink, etc. toda clase de. 3 of pers. todo, cualquiera βασιλεύς. 4 of abstr. absoluto, perfecto φιλότης. 5 with adj. ὅστις cualquiera que.
III: adverbial phrases and uses in n. 1 τὸ ἅπαν totalmente. 2 ἐξ ἅπαντος de toda cosa.

Morphology is what here establishes classification I/II/III; at a lower level, word sub-classes intervene (and the classes in II 5).

As a minimum sample I give below an adverb which is also a preposition (improper) and conjunction: ἅμα. Naturally, what will give us our key is whether or not ἅμα belongs to one of these three classes of words:

A: adv. I non temporal use 1 with verbs of movement juntamente, en compañía de θεοὶ δ' ἅμα πάντες ἕποντο. 2 with verbs of state juntos διατελεῖν. 3 uniting two nouns, adjectives or adverbs también.
II: temporal use 1 with verbs of action or process a la vez ἅμα ἔπος τε καὶ ἔργον. 2 with two verbs apenas ταῦτα τε ἅμα ἠγόρευε καὶ τὸ ἔργον προσῆγε.
B: prep. 1 with dat. 1 with verbs of movement, without temporal value junto con. 2 with temporal value al mismo tiempo ἅμα ἠῷ.
II: with gen. junto con πολλῶν ἀγοραίων.

To conclude, I should like to give some data on a preposition in very simplified form, namely on ἀπό, a preposition with only one case. There is therefore no organization of meanings based on the case governed, but only that which is habitual within the group preposition + case, that is the one obtained by combining data of the word sub-classes of the subject, the verb and the governed noun. Other data may occasionally come into play and I have already mentioned nominal transformations. For further details, I refer the reader to three articles by Angeles Martínez Valladares data may occasionally come into play and I have already mentioned nominal transformations. For further details, I refer the reader to three articles by Angeles Martínez Valladares[15].

In this case, I am not going to give a schema of the organization of the entry, a schema which, although it were as simplified as the above, would in any case be too extensive (and, at the same time, still lacking in detail). But I do wish to state that it is possible, on a purely formal basis, to give a classification of the meanings of seven large groups (from I to VII, each of them being subdivided) and all this on the strength or on the basis of the word sub-classes of the three determining elements. The translation of almost of all of these seven groups usually consists of de and other words or phrases which specify or give greater precision:

I: in the sense of distancing de, desde (not only with verbs of movement but also of process, «see», «hear», accipiendi and in transformations).
II: with value of origin de, procedente de (with verbs of origin and transformations).
III: with instrumental value a partir de, con (with verbs of process, of «manufacturing», and economic means, with adequate government).
IV: with causal values 1 a parti de, por causa de (with verbs of process and government of pers. or abstr.). 2 de, dependiente de, en virtud de (in predicative uses).
V: with indication of distance de, lejos de, a distancia de (with verbs of situation and state, with govd. nouns of pers. and place).
VI: with temporal values luego, después de (the govd. nouns indicating divisions of time or temporally fixed actions are the key).
VII: with partitive and price value de (highly specific contexts with prep. governing numerals or pronouns and verbs of «eating», «drinking», etc.).

It is remarkable in lemmata such as this one that the translation is sometimes the same in several items: the generic translation (de) is by far the most common, appearing in every one of the seven groups; some of the specific translations appear in several groups, but not so frequently.

Moreover, the formal definitions are quite clear, although they are multiform within one group: not only those of the items with Roman numbers mentioned (something of this may be seen in the preceding summary), but also of those with Arabic figures. Thus, within II 1 (values of distancing), we have constructions with subj. of person or assimilated element + verb of transit + ἀπό + place; the same, but with nouns of place or parts of the body; with verb assimilated to those of movement (ἐλευθερωθῆναι, for example) and pers. or abstr. Items 2, 3 and 4, the translation of which is in fact the same, have a similar richness.

IV

After this, I should like very briefly to define what could be considered as the contribution of the DGE to lexicographic science, and particularly to the problem of what can be expected from a bilingual dictionary, as likewise what cannot be expected.

A dictionary of this type is not only an accumulation of data brought up to date, as some people seem to think. It also entails an interpretation of these data, achieved not only by resorting to specialized instruments (translations, lexicons, commentaries, bibliography), but also by direct reading. Above all. it entails a semantic organization of the lemmata carried out from the point of view of the target language, so that the dictionary makes clear which uses of the lemma (in which context, etc.) require which translation into the target language, in this case Spanish. All this must be done by means of a staggered branching of the meanings.

Naturally, a new dictionary that aspires to more completeness than previous ones attempts to surpass them in these three respects by resorting to the data and interpretations now available to us and above all, by contributing original work in the field of semantics (for current research does not offer a great deal in this area), within the current state of the science. Yet one should not forget it has certain limits.

In the first place, this is a dictionary that keeps to realistic proportions with regard to authors and possible users. The total Thesaurus of the Greek language (and I would say of all languages) is today a utopia and would to a certain extent be uselessly repetitive. Certainly, lemmata may have not been recorded by us, or forms and meanings. These problems may be solved once Irvine's TLG is completed and made available to users, but not now with the material at our disposal. However, it is my estimate that the numbers of the new lemmata that will have to be introduced in future will not be too great.

One should not forget that a bilingual dictionary such as the DGE should necessarily be basically a handicraft work, in the old style. On the other hand, it is a work that in the social and economic conditions of a our time necessarily must be carried out by a numerous team, and which is subject to countless problems that are often more exhausting than the scientific ones proper (problems of renewing the team, financing and, unfortunately, in Spain a worse problem than these: administrative red-tap). Certain errors, certain inequalities, are perhaps inevitable: the most one can aspire to is to reduce these to a minimum. Therefore, pointing out a few errors of lacunae says nothing against either the DGE nor any other work of this kind, whenever, as I say, they are reduced to a minimum.

Neither society nor our philologist colleagues are often willing to understand a work of this kind. Good proof os the foregoing on the lack, or practical lack, of attention to semantic problems. Another proof is two irritating questions put only too often to the author of a work such as this.

The first of these is «How long are you going to take to finish it?» It is impossible to answer; the team splits up and then once more laboriously groups, its members are pressed by other occupations, more and more new material continuously appears, the printing process take longer than was ever expected, proof-reading takes up incalculable time, economic and administrative problems at times becomes thorny and unbearable. All that I can reply is that one works at normal, constant rhythm, that work is not abandoned, that the standard of quality increases progressively, as we hope.

The other annoying question, of a very different type, is «Do you use computers?» Of course we use computers to the extent that they furnish new material or else we use the material produced by those who have used them before us. Every scholar should profit from this new aid. But computers furnish material without lemmatizing it properly, nor do they organize the data of one lemma semantically or select the right contexts for organizing same. They are not the deus ex machina of modern lexicography except at he level of indices and concordances (nota bene, often inferior to those already in existence). These is a great deal of ignorance on this matter. The lexicographer is still either one man or a team of men, and computers are just one of the many suppliers of material and no more.

I return to the subject at hand: the limitations of a dictionary. One has to decide which authors and texts it will refer to (from which date to which date) and what kind of words, within these, it will refer too. Above all (and in this I vigorously defend my stance), it must supply data which the reader can check and not mere speculations of the authors as to which is the true original text for such and such a word, or whether this or that work is spurious or not. A dictionary such as the DGE should offer a system of references; whoever uses it should look up the place quoted in the edition quoted and he will at time find further information there. Whenever we cite important variants that are alien to previously established editions, this is stated. The DGE refers to a highly specific corpus of texts.

The phase of interpretation is naturally more subjective as is that of semantic classification. There will be agreement or disagreement on these points with other authors, according to case. Moreover, the necessary limitations imposed by time and space prevent one from quoting the bibliography consulted for each translation or interpretation and prevent one even more from criticizing editions that were not followed. The most essential points is that the basic data should be objective, as mentioned above.

Lastly, there is the problem which is the most important for us, the one repeatedly mentioned with regard to the semantic organization of lemmata. The first thing that might be asked (and which is asked) is why we have chosen Spanish as a target language and not English. There are obvious reasons in our case, of course. Others may choose English or any other language. What should be said to this respect, for not everybody is aware of the fact, is that the organization of the lemmata in a Greek Spanish Dictionary is often very different to what it would be in a Greek-English one. Meanings are established on the strength of a target language and, in the same way as Greek and Spanish do not possess an isomorphous lexicon, that is, one which consists of words that are semantically totally equivalent to each other, the same occurs with Spanish and English. This is unfortunate, but it is so. Others could make a Greek-English dictionary with our material, although this would mean re-organizing countless lemmata, choosing keys or contexts which define the meanings of English (sometimes, they would be the same as in Spanish, sometimes not). We could not compile this dictionary although we give in any case a model of how one should proceed in our opinion. For the rest, if we contribute our grain of sand to breaking the ever-increasing monopoly of English in the scientific field, it will not be too bad a result. Perhaps classical philologists will thank us for helping them to enlarge their linguistic experience with a language like Spanish, which they will not find too difficult.

With regard to the possible novelties of our semantics, I do not whish to over-insist on this point. What I should like to point out is that our fixing of contexts to be later organized in semantic units at several levels does at times surpass the needs of translation. One and the same Spanish translation (at times enlarged upon by means of semisynonyms) is to be found in several items. At other times, the same translation, despite its being quite generic, is specified by means of data from general contexts.

This means that we think that a dictionary thus conceived, apart from its purely lexicographic value, has syntactic interest, too: it offers a description of the diverse constructions of a lemma. It also has morphological interest: we also add data on the connection between morphology and semantics. But above all, it has a cultural interest: our treatment of words such as ἀγαθός, ἁγνός, ἀλήθεια, αἰδώς, in vol. I, for example, is important, to our mind, from several points of view (religious, social, political, scientific). An extensive dictionary should go into depth not only in the knowledge of the lexicon of the source language quantitatively, but also in that of its connections to syntax, morphology, word sub-classes, the cultural environment and Science as a whole. The limit between the mere traditional dictionary. On the basis of the points of view expounded here, we have to a certain extent penetrated to the level I have attempted to describe. Of course, it is possible to go further, turning certain items into monographs on either syntax or certain concepts of Greek culture.

A work such as the one I refer to is for ever at a stage of testing and sounding out. It cannot aspire to perfection and not only does it not shun rectification, but it welcomes same (I have already mentioned the Supplements which we are publishing). In any case, it aspires to opening up a dialogue that will break the isolation between linguists and classical philologists, lexicographers and semanticists, and between scholars of diverse languages. The aim of this article was to do the utmost to enable this to come about.

Notes

[1]