BMCR 2022.02.39

The Cambridge Greek lexicon

, , , , , , The Cambridge Greek lexicon. Cambridge; New York: Cambridge University Press, 2021. Pp. 1529 in 2 volumes. ISBN 9780521826808. $84.99.

The publication of The Cambridge Greek Lexicon (henceforth CGL) has been eagerly awaited. First conceived of by the late John Chadwick in 1997, CGL has been brought to fruition by an editorial team based in the University of Cambridge, consisting of James Diggle (acting as editor-in-chief), Bruce Fraser, Patrick James, Oliver Simkin, Anne Thompson, and Simon Westripp (p. vii of the Preface briefly discusses the history of the entire project and details the involvement of each of the editors). The back cover (see also Preface, vii – viii, and the publisher’s blurb) explains that CGL aims ‘primarily to meet the needs of modern students, but is also designed to be of interest to scholars’. Thus, this is not a full-length dictionary intended to replace LSJ, or indeed Montanari’s The Brill Dictionary of Ancient Greek (which is not mentioned in the Preface), the two large-scale Greek–English dictionaries currently on the market.[1] The editors have from the start set their sights on a different target: CGL was originally conceived of as a revision of the abridged version of LSJ (the ‘Middle Liddell’), until it became apparent that this template was too antiquated and the decision was made to compile a new and independent lexicon.[2] CGL should now certainly replace the ‘Middle Liddell’ on the bookshelves of students and scholars alike. Yet comparison with the full-length lexica is inevitable, given that CGL has turned into a much more substantial work than the ‘Middle Liddell’ and that it has been so long in the making. Furthermore, it boasts being ‘based upon principles differing from those of existing Greek lexica’ (including LSJ and Montanari), with entries being organised ‘according to meaning, with a view to showing the developing senses of words and the relationships between those senses’ (back cover). Finally, the editors of CGL aim for up-to-date coverage, as they ‘have systematically re-examined the source material (including that which has been discovered since the end of the nineteenth century) and have made use of the most recent textual and philological scholarship’.

The two main defining features that set CGL apart from other lexica, then, are its scope and the way it organises its entries. These two features determine the structure of the present review: we first comment on issues broadly related to scope, and then turn to some of the consequences of the decision to base CGL on the criterion of ‘meaning’. In both cases, there are positive and more critical matters to report.

The relatively moderate scope of CGL brings with it a number of advantages, not the least of which is that it allows for a refreshing clarity in presentation. On opening one of the volumes (the work had to be printed in two volumes to make the books easy to handle; but since they are so handsomely produced and, moreover, come with beautiful dust-jackets and a case, there is really little reason to complain) one is immediately struck by the accessible lay-out. Where the more exhaustive lexica like LSJ and Montanari look cluttered owing to their scope and the unsparing inclusion of quotations, CGL is a model of clarity, allowing each heading within its entries to begin on a new line, so that readers do not have to plough through endless scriptio continua to find the information they are looking for. Typographically, the information is clearly differentiated. Head-words and translations are printed in bold (as in Montanari), which makes sense as this is generally the most pertinent information. Translations are often preceded by more detailed descriptions of meaning, printed in regular typeface (which certainly helps to bring out ‘the developing senses of words and the relationships between those senses’ [Preface, vii], as we will discuss in more detail below): γνώμη 1 ‘faculty or application of thought, thought, thinking, judgement, understanding’. Information concerning usage is generally put between brackets: γνήσιος 6 ‘(of a literary work, a letter) genuine, authentic’. Small caps are used for grammatical usages, ‘w.gen.’, ‘pass.’, etc., author and genre abbreviations are printed in a font without serifs, and italics are used for a number of purposes.

The headings richly provide the reader with information concerning usage. Consider αἰσχρός 5: ‘(of actions, circumstances, speech, or sim., as a term for general disapproval or opprobrium) shameful, disgraceful (freq. opp. καλός excellent, admirable).’ Such elaborate descriptions concerning the ‘typical circumstances in which a word may be used’ do indeed contribute to ‘giving fresh insights into aspects of Greek language and culture’, as promised on the back cover. However, the way the information is presented is sometimes awkward. For example, the first meaning given under ἀνάσσω is ‘(of deities, deified heroes, kings) be ruler, hold power’. The second heading begins as follows: ‘(wkr.sens., of Aiolos) have control over, govern’. This would seem to suggest that this second meaning is confined to the god Aiolos, but as we read on we see that it extends to Apollo and Persephone. The designations ‘wkr.sens.’ and ‘of Aiolos’ should have been clearly differentiated to show that they apply to different levels of the heading. Similarly, ἀραρίσκω 6 begins as follows: ‘(intr., of a nightingale, as an exemplar of grief) be fixed – w.dbl.acc. in someone, in the heart (i.e. in someone’s heart) S’. However, the intransitive use turns out not to be confined to the very specific example the editors have in mind here (Sophocles, Electra 148 ἀλλ’ ἐμέ γ’ ἁ στονόεσσ’ ἄραρεν φρένας), but can also be used with reference to a festival: ‘have a secure outcome’.[3] This example also illustrates how sometimes too much information is given. If the reader looks up ἄραρεν at Electra 148, they do not need to have all the contextual information (the subject, its metaphorical connotations, the construction, the explanation of the construction) available in the text repeated in the dictionary. On the other hand, for someone reading the entry for its own sake this mode of presentation exaggerates the relevance of a rather singular phenomenon.

In order to allow for such ‘full descriptions of meanings’ and ‘illustration of usage in a wide range of passages’ (to paraphrase the Preface, viii), Greek quotations have been largely sacrificed and citations have been limited to author names or genres. In general, the benefits of this decision arguably outweigh the drawbacks. In a few cases, however, the lack of a Greek quotation or paraphrase is keenly felt. For example, a reader confronted with an idiom like οὐκ ἔχω τίλέγω finds the relevant information under ἔχω 26: ‘(intr., in neg.phrs.) have knowledge, know – w.indir.q. what one is to do, or sim.’ This mode of presentation does not seem the most effective way of helping the reader find this usage. Also, where Greek phrases are given to illustrate particular senses, it is done a bit haphazardly; it is striking, for example, that the three uses of μέντοι distinguished by CGL are all illustrated with quotations, while the three uses of καίτοι are not. With respect to citations of specific passages, the editors argue that they ‘can be unhelpful to the learner, and, by their selectivity, are in danger of giving a partial or distorted picture’ (Preface, viii). This point is well taken; however, it can be countered that in the presentation of CGL itself (as in the case of ἀραρίσκω/ἄραρεν mentioned above) it is also often not clear whether a particular usage that is said to occur in an author represents a hapax, a more or less special idiom, or a truly typical sense of the word.

Beside its clarity in presentation, CGL is very generous in its inclusion of entries for the principal parts of irregular verbs and other more or less irregular forms, consisting of a cross-reference to what used to be called the ‘dictionary form’. Not only usual suspects like εἶδον or πρήσσω get such a cross-referring entry (in the case of εἶδον to aorist infinitive ἰδεῖν, which CGL prefers as a head-word to LSJ’s and Montanari’s non-existent *εἴδω), but also, for instance, pseudo-sigmatic aorist ἔμεινα, irregularly augmented imperfect εἱπόμην, and all three current genitive forms of ἐγώ (μου, μεο and μευ, as well as ἐμοῦ, ἐμέο and ἐμεῦ). This is likely to irritate any remaining Crocker-Harrises in schools and at universities, but it must be feared that it is increasingly necessary to meet ‘the needs of modern students’ in this way.[4]

After considering the advantages of the limited scope of CGL it is time to face the costs. In principle, CGL’s coverage ‘extends from Homer to the early second century AD (ending with Plutarch’s Lives)’, and includes ‘[m]ost of the major authors who fall within that period’ (vii). Both that early cut-off point and the editors’ view of who are the ‘major authors’ evince a decidedly narrow view of the canon, and that is in our view a pity in the case of a dictionary that appears at a time when the canon is well and truly being blown open. For instance, the enormous interest in the Second Sophistic and the Greek novel of the past decades (fuelled, it may be added, not least by Cambridge scholars) has not persuaded the editors of CGL to include the relevant authors (whether or not their ἀκμή falls before the first part of the second century AD), even though that interest is evidently trickling down into undergraduate and graduate courses (witness also recent ‘Green and Yellow’ commentaries on, for example, Lucian, Longus and Achilles Tatius).[5] A further omission that those who favour an expanded canon will deplore is that of Jewish authors: no Septuagint (despite its being a major source of koinè Greek), no Philo, no Josephus, no Paul; of the New Testament only the Gospels and Acts have been included. On the other end of the chronological spectrum, the First Sophists join their later counterparts in being bypassed (the coverage of presocratic philosophy is limited to Heraclitus, Parmenides, Empedocles, and Democritus as quoted in the collection of Kirk–Raven–Schofield),[6] and the Hippocratic corpus has also been omitted in its entirety (despite its being a major source of Ionic Greek). There is, then, a bias towards the old ‘classics’ and towards poetry. Later prose in particular is not very well served: there is no coverage, for instance, of Diodorus Siculus, Dionysius of Halicarnassus (in his guise as a historian or a critic) or, more surprisingly still, Plutarch’s Moralia (only the Lives have been considered). Polybius has been taken into account, but with the weird exclusion of Book 34 (perhaps the editors judged that that geographical book would clutter the lexicon with arcane terms and names, but these might have been more selectively excluded).

Now that we have considered the issue of scope, we arrive at the other main feature that sets CGL apart from its competitors: the principle of arranging entries ‘according to meaning, with a view to showing the developing senses of words and the relationships between those senses’. In general it may be argued that adhering to a single criterion throughout, instead of sometimes referring to chronology, at other times to grammar (as is usual practice in lexica), makes the overall presentation more consistent. This enhances the accessibility of the book in the sense that the readers know what to expect when they come upon an entry. At the same time, such an approach does not always do justice to the great variety in semantic structure exhibited by different (kinds of) words, and this defect comes at the cost of user-friendliness, as we will explain below. (Assessing the merits of this meaning-oriented approach is made more difficult by the fact that the Preface, rather surprisingly, has almost nothing to say about the editors’ views on the nature of semantic structure and semantic development. However, Diggle has on several occasions expounded the principles underlying CGL in longer talks and lectures, which owing to the Covid pandemic were mostly held online and are still easily accessible for those interested: here are YouTube videos for Cambridge Classics, The Cambridge Centre for Greek Studies, and the Hellenic Society.)

At their best, CGL entries display a sophisticated sense of semantic differences, and the result is that such entries make for interesting reading in themselves, constituting almost mini-essays on the semantics of each word. For example, under γνώμη, we move from ‘faculty or application of thought’ to ‘result of thinking’ to ‘expressed opinion’, et cetera. Thus the reader is smoothly guided through progressively more concrete shades of meaning. Another entry with elegant and precise descriptions of meaning is αἰδώς: from ‘sense of shame (inhibiting dishonourable behaviour)’ to ‘sensitivity to the feelings, status or claims of others’ to ‘sense of propriety and delicacy (constraining self-assertion)’, etc. In the case of particularly complex entries where meanings proliferate, helpful summaries are provided at the beginning, providing further guidance. These are not simply abbreviated versions of the entry, such as appear in Montanari, but descriptions of categories of meaning. For example, under ἔχω we find: ‘(1 – 3) hold by physical contact, (4) hold in one’s possession, (5 – 7) possess as one’s own’, etc. The inclusion of such summaries is most welcome. CGL is also sensitive to the difference between meaning and metaphor. For example, the word ἄδυτος as a singular noun is (quite rightly) given one sense: ‘inner sanctum’. At Plato, Theaetetus 162a, Protagoras’ Truth is said to utter words ‘from the inner sanctum of the book’ (ἐκ τοῦ ἀδύτου τῆς βίβλου), which is a metaphor intended to mock the sophist as a would-be oracle. This is explained in CGL: ‘(fig., w.gen. of a book, fr. which the teachings of Pythagoras [sic !] spoke forth)’. LSJ follows a similar procedure; Montanari, on the other hand, here and elsewhere assumes an ‘extension’ of the regular meaning (in this case giving the translation ‘nucleus’), which is misleading.

Here we may also point out that CGL squarely meets the needs of modern students when it comes to the translations themselves. No longer do we find the shibboleths by which generations of the classically educated recognise each other; for example, ἐυκνήμις is not rendered by the awkward, hyphenated ‘well-greaved’ (LSJ), but by the more straightforward ‘equipped with fine greaves’ (still, the ‘greaves’ have stayed, and ‘shinpads’ may indeed be a bridge too far). CGL also avoids euphemism. Thus βινέω is translated as ‘fuck’ (LSJ famously resort to Latin, ‘inire, coïre, of illicit intercourse’; Montanari opts for ‘to have intercourse with’, adding the label ‘vulg.’ to pin down the register). The CGL entry matter-of-factly continues to specify that the active and passive can both be used of men and women, whereas both contenders betray subtle biases: LSJ states that the passive is used ‘of the woman’, which is more in line with Victorian prudery than Greek usage, while Montanari explicates that the passive means ‘to be screwed, also of men’ – one hopes that that ‘also’ is intended as an implicit correction of LSJ rather than to register shock or surprise.

As with CGL’s scope, its arrangement according to ‘meaning’ has downsides as well as upsides. One area in which the treatment of ‘meaning’ is somewhat disappointing is that of particles, where CGL continues to rely on rather vague Dennistonian terminology and stopgap translations. This reflects the fact that CGL is not always as linguistically well-informed as it could have been.[7] While it is to some extent understandable that not more use has been made of the literature on Greek linguistics of recent decades, the publication of which largely coincided with the work on CGL, there are missed opportunities. For example, μέντοι is said to have a ‘progressive or connective’ use, meaning ‘well, now, then’. This does not tell us anything about what the specific contribution is of μέντοι in such cases as compared to other ‘progressive or connective’ particles. Sometimes a summary or general description is given, which is most welcome, but a more discursive description of the particular meanings would also have been helpful in the example just cited. Moreover, these descriptions can be vague as well. For example, τοίνυν is described as follows: ‘The conj. [sic] has a colloq. tone, giving a weak logical connection to the previous sentence’ (before translations like so, then, therefore and well then are suggested). It is unclear what that ‘weak logical connection’ consists in, and actually the function of the particle can be readily described in discourse-linguistic terms. We would offer something to the following effect: Like οὖν, the particle signals an ascent to a higher strand of discourse (for example, when the speaker transitions to a new section of discourse, returns to the narrative main line after a digression, or gets to the point after a preparatory argument) with the added value of an appeal to the audience’s attention (marked by τοι). Cumbersome? Perhaps – but a description focused on a particle’s function rather than on translation does justice to the current state of the art in analysing this famous and important class of Greek words.[8]

But there are more fundamental issues with CGL’s approach. While the organisation of entries according to ‘meaning’ often results in a refreshingly new look, the representation of semantic structure is severely limited by the lack of hierarchy in the structure of entries. Each new meaning usually gets its own heading, so that entries are essentially lists of meanings (there are exceptions, such as in the case of the prepositions, on which see below), but meanings often do not develop in linear fashion. Consequently, the relationships between different senses are not always clear. Let us illustrate this with some examples. The entry εὔχομαι has six headings: 1. ‘speak confidently’ (etc.), 2. ‘speak boastfully’, 3. ‘speak triumphantly’, 4. ‘utter a prayer’, 5. ‘pray for’, 6. ‘promise as repayment for an answered prayer’. It will be noted that senses 1 – 3 are highly related and distinct from senses 4 – 6. However, there is no marked transition between 3 and 4. This could be taken to suggest that the meaning ‘utter a prayer’ (4) is somehow derived from ‘speak triumphantly’ (3), which surely is not intended. Is ‘utter a prayer’ perhaps regarded as being derived from ‘speak confidently’, the basic meaning of the first group? Or do both meanings derive from another concept? The reader has no way of telling. A more hierarchical structure (such as the one found in LSJ) could have made this clear.

We observe similar issues when we return to αἰδώς. After the three senses mentioned earlier, we find 4 ‘modesty, shyness etc. ’; this is followed by 5 ‘(concr.) area of the body normally concealed through modesty; private parts’, and then 6 forgiveness. What is the sense of this ordering, and how does it help users quickly to find what they need? Surely 6 does not somehow ‘derive’ from 5, which may be suggested by the present structure. If 5 is to be mentally subsumed under 4 (the former being a concretisation of the latter, and so in fact forming an entry 4(b)), does 6 somehow derive from 4 (‘forgiveness’ from ‘modesty’)? Or are the main senses listed under 1 – 4 and 6 variations on the same general concept (described at the beginning of the entry) and do none of these depend on one another? After 6, we move on to 7 ‘feeling of embarrassed shame (over one’s present situation); shame’ and 8 ‘feeling of guilty shame (for past actions); shame’. These highly related senses might easily have been subsumed under a single heading. Moreover, it is once again implausible that we should regard these senses as developing from the previous sense (‘forgiveness’). In a more hierarchical presentation a clearer distinction could have been made, for example, between αἰδώς as an inhibitory force and αἰδώς as a feeling of shame, with more specific senses listed under each heading.  The inclusion of summaries with some entries (see above) only partly mitigates the problem.

The second main issue that we wish to highlight is that the focus on ‘meaning’ tends to come at the cost of due consideration of morphosyntactic criteria. We will illustrate the point with reference to three of these: case, voice, and verbal complementation.

It hardly needs pointing out that the meaning of a word is often affected by the particular case form it governs. The most obvious example is constituted by the prepositions. Readers confronted with a preposition used in a sense unknown to them will surely take the case form of the following noun phrase as their starting point of enquiry. However, CGL does not organise entries of prepositions according to case, but according to domain of experience (location, movement, time, etc.). Even if readers manage to identify the correct domain, the subheadings within these domains are still not organised by case but by ‘meaning’; however, where the relationship between meanings can be made clear in the case of content words, this is much less perspicacious here. For example, under ἐπί A (‘location or space’) we find:

1. on the top of, on, upon – w.gen. or dat. the earth or ground, an object, part of the body, or sim. Hom. + (as adv.) on top, thereupon Hom. Hdt.
2. within the limits of, on – w.gen. or dat. an island, piece of land Il. +
3. over – w.acc. an area, part of the body, vel sim. Hom. +
4. as far as w.acc. a limit or distance Hom. +
5. at – w.gen. or dat. a location Il. + [this goes on to 15].

In meanings 1, 2 and 5 the preposition governs a genitive or dative, but in meanings 3 and 4 it governs an accusative. The separation of these syntactically related categories is presumably due to the supposed separateness of their meanings (3 ‘over’ is perhaps more akin to 1, 2 ‘on’ than 5 ‘at’ is) but whether this mode of presentation is helpful to the reader is debatable (and given the highly idiosyncratic use that languages make of prepositions, the question whether ἐπί ‘means’ ‘over’ or ‘at’ or whatever is more a question of English idiom than of Greek). Strangely, the meaning ‘on’ (with genitive) then returns in headings 12 – 15, which are separated from the rest because here more specific information is given concerning usage (‘ref. to sitting’, ‘ref. to being conveyed’, etc.). This, again, makes the structure inconsistent.

Similarly, with certain verbs the different meanings are mostly divided among categories of voice (albeit with some overlap). A reader confronted with a certain voice form in the text will find themselves frustrated if they use this criterion as their starting point for a search in CGL. For example, ἄρχω has twelve headings. Two of these (2 and 8) begin with the information ‘mid.’, but these uses are intermingled with the other uses. It is true that we find certain meanings both in the active and middle, but even in those cases the middle may be used in a different way from the active. For example, under 6 ‘start (fr. someone, sthg., somewhere); begin’, we find, under ‘mid.’, ‘w. Zeus (in making a libation)’, a use that seems exclusive to the middle. It would have been useful to have separate active and middle headings so the reader could see which uses are exclusive to the middle, even if this would have resulted in some structural redundancy. (We do applaud the exclusion of redundant ‘passive’ headings – of the type ‘loosen’; pass. ‘be loosened’ – that are found in Montanari.)

This issue is even more pressing for verbs that have clearly distinct transitive and intransitive uses. For certain major verbs, CGL does make this distinction a guiding principle, so that ἵστημι is clearly separated from ἵσταμαι and φύω from φύομαι. In other cases, CGL departs from this procedure. In the case of ἀραρίσκω, this yields a strange hybrid. The word is first given 11 senses. Then follow subentries for ἀραρότως, ἀραρώς, ἄρμενος and ἄρμενα (the fourth of which might have been subsumed under the third). Now, we find intransitive uses scattered throughout the first 11 headings. For example, heading 5 gives the meaning fit, fix; this is followed by ‘(intr., of a person’s arm) be fixed securely’. Then follow headings 6 and 7, which both begin with the designation ‘intr.’. With heading 8 we return to a transitive usage (equip, furnish). To our minds, this is in itself convoluted, but what is worse, the reader may have no idea that these intransitive senses are limited to certain morphological categories – for example, the instances listed under heading 6 both feature the perfect. This stands in stark contrast with the use of subentries for specific forms, as noted above. In LSJ and Montanari, these forms can be simply subsumed under the broader heading of intransitive meanings, and this is surely more elegant and user-friendly.

A final area in which neglecting morphosyntactic issues has in our view had detrimental consequences is that of verbal complementation. The sense of verbs is often co-determined by the form of the complement they take. Take for instance ὁράω. When that verb expresses direct sensory perception (as in English ‘Jane saw John walk past’), it is complemented by an accusative-and-participle (almost always in the present); when it expresses cognitive understanding (compare English, ‘I now see that you were right’, or ‘I do not see who is right’), it is complemented by an accusative-and-participle (in all tenses), a ὅτι/ὡς-clause, or an indirect question. Finally, ὁράω can also be followed by a so-called ‘effort clause’ (usually ὅπως + future indicative) and then means something like ‘see to it that’.[9] Here are the relevant parts of CGL, s.v. ὁράω:

4 discern with the eyes, see, perceivepersons, things (sts. doing sthg., or in a certain state) Hom. + (…)
5 perceive with the mind, become aware, understand, realise – w.indir.q. or what (or that sthg.) is the caseHom. + – w.nom.ptcpl. that one is acting wrongly or sim. E. Th. – w.acc. + inf. that sthg. is impossible Th. (…)
7 (…) see to it, take care – that sthg. shd. (or shd. not) be the case Hdt. S. E. Th. +

The divisions in meaning that are made here are sensible, but the information about constructions leaves much to be desired. Under 4 (ὁράω as a verb of direct sensory perception), the regular construction with the (present) participle is not explicitly given, but can only be inferred from English ‘doing something’. Under 5 (ὁράω as a verb of understanding), the only constructions given initially are indirect questions and complement clauses, while the accusative-and-participle is left out. Under 7, we again find a general reference to a ‘complement clause’, with nothing to differentiate the declarative complement clauses (with ὅτι/ὡς) intended under 5 from the ‘effort clauses’ (with ὅπως) intended here. With so little guidance on constructions, users are deprived of a useful instrument when they need to make up their mind about what a given instance of ὁράω might mean.

The rest of the entry under 5 is also more confusing than illuminating. First, it is stated that the verb sometimes occurs with a so-called nominative-and-participle construction and then means ‘(understand) that one is acting wrongly or similarly’. The references must be to E. Med. 350 καὶ νῦν ὁρῶ μὲν ἐξαμαρτάνων (‘I now see/understand that I am making a mistake’), and Thuc. 7.47.1 τοῖς τε γὰρ ἐπιχειρήμασιν ἑώρων οὐ κατορθοῦντες καὶ τοὺς στρατιώτας ἀχθομένους τῇ μονῇ (‘for they saw that they were not succeeding in their undertaking, and that the soldiers were finding their stay burdensome’). It is questionable, however, that we here have to do with a separate construction and a separate sense of the verb: it is merely the case that the subject of the main verb ὁράω and of the complementary participle are identical, and in such a case we find a nominative rather than an accusative participle (in the Thucydidean passage the nominative-and-participle is indeed coordinated with an accusative-and-participle – without, of course, a change of meaning of ὁράω, which expresses cognitive understanding in both cases).[10] The second addition to the basic information given in 5 is also oddly specific and in that sense misleading. The statement that ὁράω can be followed by an accusative-and-infinitive and then means ‘(understand) that something is impossible’ is nothing more than an interpretation of the single passage in classical Greek in which ὁράω appears to be followed by an infinitive (Thuc. 8.60.3: ἑώρων οὐκέτι ἄνευ ναυμαχίας οἷόν τε εἶναι ἐς τὴν Χίον βοηθῆσαι, ‘they realised that it was no longer possible to come to the rescue of Chios without a sea battle’). Even if one keeps the infinitive (some editions bracket it), the presentation makes it seem as if the entry at this point introduces a regular meaning and construction of ὁράω (remember the Preface’s promise to specify ‘the typical circumstances in which a word may be used’). In cases like these, the fact that CGL doggedly sticks to its principle of not citing places becomes a problem, and we return to a recurrent issue: instead of giving fundamental meanings (basic senses of a word), the lexicon here becomes a commentary on specific (and clearly not representative) passages, providing context-based translations. Perhaps there is nothing wrong with that, but it would have been helpful if the user had been given the means to differentiate between the different sorts of tasks (translation-help and lexicographical description) CGL performs.

All in all, CGL is a pleasure to use due to its attractive presentation (both on the outside and on the inside), its contemporary translations, the wealth of information it provides concerning usage, and the fresh insight it provides in the semantic structure of certain key concepts in Greek culture. It will no doubt become a standard work of reference for students throughout the world and it will save scholars a trip to the fuller LSJ or Montanari for most quotidian purposes. However, some unfortunate decisions in coverage, together with its somewhat experimental approach to lexicography, make it a flawed colossus.


[1] H.G. Liddell, R. Scott, and H.S. Jones, A Greek-English Lexicon, 9th ed. Oxford 1940 (and many reprints); F. Montanari, The Brill Dictionary of Ancient Greek, Leiden 2015. It is reviewed in BMCR 2018.03.46.

[2] H.G. Liddell and R. Scott, An Intermediate Greek Lexicon, Oxford 1889 (and many reprints).

[3] One detects the trace of the LSJ entry, which opts for a completely different translation: ‘please, gratify, “ἐμέ γ᾽ ἁ στονόεσσ᾽ ἄραρεν φρένας” S.El.147 (lyr.); ἁ Νεμέα ἄραρε Nemea smiled on [him], Pi.N.5.44.)

[4] Still, one sometimes wonders. Do we really need, as successive entries: ‘εἴδομεν1 ( see ἰδεῖν; εἴδομεν2( see οἶδα; εἰδόμην (aor.2 mid.): see ἰδεῖν; εἶδον (aor. 2): see ἰδεῖν’?

[5] See, N. Hopkinson, Lucian. Selected Works, Cambridge 2008; E. Bowie, Longus. Daphnis and Chloe, Cambridge 2019; T. Whitmarsh, Achilles Tatius. Leucippe and Clitophon: Books I–II, Cambridge 2020.

[6] G.S. Kirk, J.E. Raven, and M. Schofield, The Presocratic Philosophers, 2nd ed., Cambridge 1983.

[7] A minor point that may be mentioned in this context is the antiquated reference to different types of aorist by ordinal numbers (first aorist, second aorist) instead of the more informative designations ‘(pseudo)sigmatic aorist’, ‘thematic aorist’ and ‘root aorist’.

[8] On τοίνυν see C.M.J. Sicking, and J. M. van Ophuijsen, Two studies in Attic particle usage Leiden 1993, pp. 152 – 164; G.C. Wakker, ‘“Well I will now present my arguments.” Discourse cohesion marked by οὖν and τοίνυν in Lysias’, in S.C. Bakker and G. C. Wakker (eds.), Discourse cohesion in Ancient Greek, Leiden 2009, pp. 63 – 81.

[9] One may consult E. van Emde Boas, A. Rijksbaron, L. Huitink, and M. de Bakker, The Cambridge Grammar of Classical Greek, Cambridge 2019, parr. 41.3, 44.1, 52.9-10.

[10] The whole thing reads like a badly digested version of part of LSJ, s.v. ὁράω II.d: ‘c. part., καπνὸν . . . ὁρῶμεν ἀπὸ χθονὸς ἀΐσσοντα we see it rising, Od. 10.99; (…) rarely in reference to the subject, ὁρῶ μὲν ἐξαμαρτάνων (= ὅτι ἐξαμαρτάνω) E. Med. 350’.