[image ALT: Much of my site will be useless to you if you've got the images turned off!]
Bill Thayer

[image ALT: Cliccare qui per una pagina di aiuto in Italiano.]

[Link to a series of help pages]
[Link to the next level up]
[Link to my homepage]

This page has been carefully proofread
and I believe it to be free of errors.
If you find a mistake though, please let me know!

 p9  Introduction

§1 Riddle-writing before Symphosius

Man has ever liked to speak in metaphor, and in metaphor is to be sought source of riddles.1 Lindley notes2 that "Riddles play upon analogies among things perceived. Essentially the primitive mode of invention is as follows: Some one discovers a new analogy among natural objects, formulates a question concerning this, and thus a new riddle is born." Aristotle was the first to point out the close connection between riddles and metaphors:3 "While metaphor is a very frequent instrument of clever sayings, another or additional instrument is deception, as people are more clear conscious of having learnt something from their sense of surprise at the way in which the sentence ends and their soul seems to say, 'Quite true, and I had missed the point.' . . . This, too, is the result of pleasure afforded by clever riddles; they are instructive and metaphorical in their expression."

With a race as fond of metaphor, simile, and allegory as the Semite, riddles and enigmatical have always been popular. We recall the story of Samson and the lion that he killed, in whose carcass bees stored their honey, and the riddle he propounded to the Philistines:4 "Out of the eater came forth food, And out of the strong came forth sweetness." When the Queen of Sheba visited Solomon to put his reputation for wisdom to the test, a contest of wits ensued which consisted mainly in the alternative propounding of riddles and their solution.5 The Bible contains none of these, but they are preserved in the Talmud (Midrash Mishle) and the Second Targum to the Book of Esther. Of these the best known is the enigma of Lot and his daughters, a favorite throughout the Middle Ages.6 The correspondence  p10 of Solomon and Hiram in regard to the building of the Temple at Jerusalem took the form of a riddle-strife, from which Solomon emerged with the honors due a wisdom so great.7 Such riddle-contests were quite popular in the Middle Ages, cf. e.g. the anonymous Salomon et Marcolfus,8 the Salomon et Saturnus,9 the Altercatio Hadriani et Epicteti,10 the Disputatio regalis iuvenis Pippini cum Albino scholastico,11 etc.

Among the Greeks, in whose estimation mental agility was a virtue, the vogue of the riddle was widespread and persistent. There were two recognized types, the αἴνιγμα and the γρῖφος12 (both words occur in Latin as aenigma and griphus respectively). Perhaps the oldest riddle was that of the Sphinx, which Oedipus solved to that creature's discomfiture and death (for the riddle in prose cf. Apollod. III.8; in verse, Athen. X.83; Anth. Pal. XIV.64). The references to it are numerous, e.g. Hes. Theog. 326; Soph. Oed. Tyr. 391; Eurip. Phoen. 45 and Schol. on same; Sen. Oed. 92‑102; cf. Plaut. Poen. 443; Ter. Andr. 194; Plin. H. N. XXXIV.48. Homer is said to have died of pique at his inability to solve the world-old riddle of the louse propounded to him by three fisherboys — a riddle which occurs as aenig. XXX in Symphosius (q.v. with note containing citation and references). The entire Contest Hesiod and Homer is a riddle-strife of the kind mentioned above.13 Riddles and conundrums were favorite aids to after-dinner conversation. Our best ancient source is in the Deipnosophists of Athenaeus (X.69‑88), where for pages the subject of γρῖφοι is discussed and dozens are quoted from such authors as Cleobulina of Lindos, Clearchus, Pindar, Antiphanes, Alexis, Diphilus, Callisthenes, Pythagoras,  p11 Agathon, Plato, Simonides and others of lesser note. Oracles were commonly given in riddle form; the phrase 'Delphic oracle' became a synonym for ambiguity. The entire fourteenth book of the Anthologia Palatina is a collection of riddles, both arithmetical and oracular, in epigram form. Many of the oracular type are to be found in Herodotus.

The more sober-minded Romans were not possessed of quite the same fondness for enigmas as were the Greeks. There seem to have been few riddles genuinely Roman in origin. Perhaps the oldest one is one on the god Terminus, who refused to budge from his original site on the Capitolium, not even for Jupiter himself, quoted from Varro's De Sermone Latino ad Marcellum by Gellius (XII.6):

Semel minusne an bis minus sit, nescio;

An utrumque eorum, ut quondam audivi dicier,

Iovi ipsi regi noluit concedere.14

In the same passage Gellius informs us: Quae Graeci dicunt aenigmata, hoc genus quidam ex nostris veteribus scirpos appellaverunt. They gained the name scirpi (lit. 'rushes'), in all probability, from the involved pattern of plaited rushes.

Ovid represents Numa (whose reputation for wisdom made him to the Romans what Solomon was to the Hebrews) as being tested by Jupiter in three brief riddles which the king answers promptly and correctly (Fast. III.339‑342).15 In another passage he has Faunus approach Numa in a dream with a riddle to which Egeria finds the solution (the origin of the festival of the Fordicidia, Fast. IV.663‑672).16

Charms, incantations and imprecations, as well as oracles and prophecies often took an enigmatical form. A fifth century Latin writer on medicine has preserved for us several remedial  p12 charms, which are simply riddles that probably have their origin in primitive folk-lore; for like the vast majority of riddles of all types they can be paralleled in the folk-lore and literature of other races (Marcellus, De Medicamentis XXI.3 and XXVIII.16).17

In Sen. Phoen. 131‑139 we find the age-old incest or mixed-parentage riddle, which in Greece was quite naturally associated with the Oedipus story. The same type of riddle often had Lot and his daughters for its subject.18 A variation occurs in the Historia Apollonis Regis Tyri c4.

As the Roman cena developed along Hellenic lines, Greek entertainers and entertainment became more and more popular. In Gell. XVIII.2 we have our best account of the after-dinner diversions of Roman students celebrating the Saturnalia at Athens, where griphi and aenigmata of all sorts were in order.19 At Trimalchio's dinner a riddle is propounded for the guest to solve (Petron. 58), and the apophoreta, or 'favors', were distributed by means of tickets (pittacia or tesserae) on which riddles concealing the names of the presents were written20 (Petron. 56). Augustus was fond of distributing gifts in like manner titulis obscuris et ambiguis (Suet. Aug. 75).

To name-riddles the Romans contributed but little.21 One of Greek origin occurs in Ov. Remain. Amor. 475‑476. Of Greek origin, too, was the riddle on the fish (ἰχθὺς) as the symbol of Christ (August. Civ. Dei XVIII.23). In a Pompeian wall-inscription we have one involving a pun on with meanings of sui (CIL IV.1877).

Among the later writers Ausonius was especially fond of riddling in all its forms. Many of his poems are acrostics or anagrams or epigrams, the solutions of which are concealed from the reader but indicated in their wording, so that they may be easily guessed or worked out, e.g. cf. Epist. XIV.71‑81 (= liber XVIII.14.71‑81) and his Griphus Ternarii Numeri (= liber XVI) with its prose introduction, which, brief as it is, has value  p13 for us as being, along with Gell. XII.6, the only discussion of riddle-writing in Latin, for Latin literature has nothing comparable to the chapters in the tenth book of Athenaeus.

§2 Symphosius

And yet Latin literature has bequeathed to posterity the only complete collection of ancient riddles, the work of a single author, that we possess. Amid the writers of the Latin Anthology we find (Riese 286) a century of riddles under the name of Symphosius. Nothing further is known of him beyond the name, and even this has been the subject of dispute. For in the otherwise excellent edition of August Heumann (Hannover 1722) the absurd thesis was set up that these riddles were the lost Symposium of Lactantius mentioned by Jerome.22 Some later editors, accepting this opinion too uncritically, even went so far as to include hundred riddles in editions of Lactantius (e.g. Migne, Patr. Lat. VII.289‑298, and Fritzche, ed. Lactantius, Leipzig 1845, II.298‑308). This confusion of a 'Caelius Firmianus Symphosius' with Lactantius Firmianus had already been made by scribes as early as the tenth century.23 Wernsdorf24 opposed Heumann's contention on the grounds that the poet's name is mentioned in the first line of the praefatio, and that Symphosius is quoted by name by several early writers, notably Aldhelm. Paul25 rather easily disposes of Heumann's argument in similar fashion. Yet modern scholars have abandoned the name 'Caelius Firmianus Symphosius',26 approved by Pithoeus27 and Wernsdorf, for the author of our riddles, for it rests upon an exceedingly uncertain ascription of two minor poems in the Anthologia Latina to such a writer.28 And even granting the  p14 existence of such a name for the writer of the two poems referred to, there is not the slightest bit of evidence for attributing the riddles to him also. But modern editors (e.g. Baehrens and Riese) are unanimous in accepting for these enigmas an author named Symphosius.29 In this view they are supported by Paul and Schenkl.30

In regard to the date of Symphosius there has been even greater dispute than over the name. By Wernsdorf31 he was assigned to the fourth century; by Paul32 and Schenkl33 to the latter part of the fourth or the fifth; L. Müller34 would put him as early as the second or third on account of the accuracy of his prosody; while Riese35 would have him a contemporary of the compilers of the Salmasian Codes (a collection which we know, from internal evidence, was made at Carthage between 532 and 534 A.D.);36 this opinion is accepted by Hagen.37 Riese's view would make Symphosius one of that group of African poets who flourished at Carthage under the Vandal kings Thrasamund (496‑523 A.D.) and Hilderich (523‑531 A.D.), of whom Luxorius was the chief writer and possibly himself the compiler of the Latin Anthology as we have it preserved in the Salmasian Codex. But Riese's evidence for this view of late authorship is very inconclusive. Noticing that a number of names have in addition a complimentary title such as vir clarissimus, vir inlustris, scholasticus, grammaticus, etc., he has drawn the conclusion that all authors so denoted were contemporaries of the compiler, and that those whose names appear without such title, as well as all poems appearing anonymously, were of earlier date.38 While this assumption seems plausible, too much weight  p15 should not be attached to it, for our earliest copy of the Codex Salmasianus is certainly no earlier than the late seventh century, i.e. at least 150 years after its compilation at Carthage. Meantime many a title might have been added, altered, or lost in copying. And, negatively, there is nothing to show that such a title might not have been possessed before the compilation, nor even if we grant the ascription of the titles to the compiler, can it be proved that he made his ascriptions only to living contemporaries. Riese's view, I think, must be abandoned, as well as that of L. Müller. The name Symposius or Symphosius does not seem to have been found before the fourth century.39 The fact that there is not the slightest trace of Christianity in Symphosius' work and that his enigmas deal entirely with things distinctly pagan, cannot in itself be taken as an evidence of an authorship as early as the 'New Poetry' of Hadrian and the Antonines, for paganism lived on and flourished side by side with Christianity long after the official recognition of the latter.40 And in this period of transition there were many nominal Christians like Ausonius, whose writings were thoroughly pagan in character. The spirit in which Symphosius wrote, in so far as can be judged from his brief preface, his subject matter and his handling of it, make him a kindred spirit with the late fourth and early fifth century 'scholastics' such as Ausonius, Macrobius, and Martianus Capella, trained rhetoricians and grammarians, who delighted in jest, conundrum, and quip for mental recreation. I would therefore place Symphosius in the late fourth or early fifth century, and this date first advanced by Wernsdorf and supported by Paul and Schenkl is now accepted by the leading height i.e. of Latin Literature41 as well as by the Thesaurus Linguae Latinae.42

 p16  §3 Language and meter; style; subject-matter

The enigmas of Symphosius consist each of three hexameter lines. Their original number was 100, but No. XCVI has disappeared from the MSS., in some of which various attempts were made to supply the gap. They are preceded by a preface of 15 (or 17)43 hexameters, in which the author states the occasion and purpose of writing this century of riddles. The Latinity is excellent, there being virtually no departures from the classical norm. The use de with the abl. for ex with the abl. or for the gen. (praef. 2, 9, 15; L.1; LXXV.3) and dum with the imperf. sub. as a virtual equivalent of cum (praef. 3), to which due attention has been called in the commentary, can both be paralleled in classic authors and are not in themselves to be taken as too definite an indication of Symphosius' lateness.

His meter is of equal excellence.44 He admits elision only with est, and this only in six instances.45 There is but one instance of lengthening before caesura.46 There are only two cases of metrical infelicity: in one, the first syllable of profeci is arbitrarily shortened;47 in the other, the first syllable of rubida must be read long.48

The correctness of Symphosius' Latinity and prosody is quite generally conceded and commended, but not enough attention has been paid to his style. To one who reads through the 315 hexameters with an ear attentive to rime, assonance, and alliteration the author's extraordinary cleverness is at once apparent.49 It is perhaps in his fondness for these devices that he gives the greatest evidence of his kinship with the writers of the fourth and fifth centuries; of his skill in their use there can be no doubt.  p17 M. Manitius50 has noted 34 instances of rime between the thesis of the third foot and the final syllable of the sixth foot,51 which in many cases is strengthened, much in the manner of the medieval Leonine, by the additional riming of the immediately preceding syllable. Some of the more striking examples are IX.3: excepit . . . recepit; XXX.1: nostrarum . . . ferarum; LXXXVIII.3: sudori . . . labori; in LXX.1 and 2 the rime is even tripled by an end-rime with v. 2; dicendi . . . tacendi . . . , loquendi. More intricate is the rime-scheme of XX, where there are end-rimes on all three lines (so . . . to . . . to) and, in addition, internal rime in the third foot of lines 1 and 2 (to . . . io). No. XXX has, in addition to the Leonine above-mentioned, end-rimes for the half-lines of lines 2 and 3 in the following pattern: capias . . . recuses, . . . capias . . . reportes. No. LXXXII exhibits a similar arrangement in lines 1 and 2: fuimus . . . uno . . . , unus . . . uno. No. XXI shows the internal rime only: facies . . . dies; cf. XCII.1 and 2. Manitius has made an additional classification of rimes between the theses of the second and fourth feet, between the thesis of the fourth foot and the final syllable, and of such internal rimes within the individual lines (other than those of the first class above-mentioned) he finds 55 examples. The last of these types is especially frequent, owing to a combination particularly favored by Symphosius for the second half of his lines, e.g. No. II.1: ripae vicina profundae (observe also 2 and 3); V.1: multos habitura ligatos; IX.1: longa delapsa ruina; XIII.1: formosae filia silvae, etc.52 As an example of assonance not involving rime might be adduced the succession of u‑sounds in LIV.1: Exiguum munus flexu mucronis adunci; and a still more remarkable double assonance (u and qu) in LXXIX.3: Ducor ubique sequens et me quoque cuncta sequuntur.

 p18  More striking by far than his tendency toward rime is his mastery of alliteration and word-play. Frequently the two are joined, as in V.2: Vincior ipsa prius, sed vincio victa vicissim; XXIV.1: Exiguum corpus, sed cor mihi corpore maius. Sometimes this combination exists not merely in a single line but is carried on throughout the entire three verses of the enigma, e.g. No. XLIII: Pendeo . . . Pendens . . . Pendula; and the truly intricate phrasing of No. XLIV: Mordeo mordentes ultro non mordeo quemquam; | sed sunt mordentem multi mordere parati. The best example of balanced rime, assonance, alliteration and word-play all in one is No. LXX:

Les bona dicendi, lex sum quoque dura tacendi,

Ius avida linguae, finis sine fine loquendi,

Ipse fluens, dum verba fluunt, ut lingua quiescat.53

Of alliteration alone examples are to be found in almost every verse. Some lines show a clever use of double and even triple alliteration; cf. the subtle sounds of f, t, and p in No. XI.3: Flumina facturus totas prius occupo terras; of l and s in No. LXXII.1: Truncum terra tegit, latitant in caespite lymphae; of f and s in No. LXXXV.3: Pro pedibus caput est: nam cetera corpore non sunt. Sometimes there is alliteration between the first word of each line, e.g. No. LXXIII: Non . . . Nam . . . Nunc;54 cf. also LIII: Nolo . . . Nolo . . . Nolo and XLIII (v. sup.). Often single phrases of great effectiveness catch both eye and ear, e.g. No. XXVII.1: Vivo novem vitas; XLVII.i: flamma fumoque fatigor; L.3: Mole premor propria; LIX.3: Meque manus mittunt manibusque; LXII.2: Viscera tota tument; LXVII.2: intra mea membra meantes; or entire lines upon a single sound, e.g. No. LIX.1: Non sum cincta comis et non sum compta capillis; LXXIII.3: Nuncque mihi magna est animae, nunc nulla facultas.

 p19  The very nature of riddle-writing calls for word-play, and in this Symphosius abounds. In addition to the instances cited above under alliteration might be noted examples of punning; cf. No. XXXIII.3 (on tollere); LXII.2 (on robur); LXIX.3 (on ante); LXXIII.3 (on animae); LXXIX.1 (on mundus); LXXXV.3 (on sapientia); XCII.i (on sustinui) and 3 (on peregit); XCIII.3 (on pes); the punning use of the words pes and caput in several enigmas, e.g. No. XL, XLI, LVII, LXXXVI, LXXXVII; a phrase such as LVII.2: capitis vestigia = 'head-prints'; lines such as XXXIII.1: Dentibus insanis ego sum, qui vinco bidentes; XXXVI.3: Nomine numen habens, si littera prima periret. Other examples of word-play are found in the use of the same word in two forms, e.g. No. XXXVIII.1: A fluvio dicor, fluvius vel dicitur ex me; LXXIX.3: Ducor ubique sequens et me quoque cuncta sequuntur; or in such pairing off of word does as VIII.3: lucem . . . lumen; XXXIX.3: Et vehor et gradior; or in lines such as XXXIV.3: Et fera sum sapiens, sapiens fera si qua vocatur; LXXX.3: Non resono positus, motus quoque saepe resulto. But by far his most effective device is a third-line word-play on compounds of the same verb, e.g. No. IX.3: excepit . . . recepit; XXII.3: gero . . . congero; LVI.3: Dedita . . . condita; LIX.3: mittunt . . . remittor. Some of these many word-plays can be rendered, or at least approximated, in English; others are beyond the reach of the translator.

Symphosius has but few borrowings from earlier authors that can be identified with any degree of certainty. Undoubtedly Vergilian, however, are XIII.2; LVI.1; LXXVII.2; XCV.1. No. XLII.3 is undoubtedly from Horace, though the brief phrase is used in a totally different connection. In LX.2, Symphosius uses the word frondicomus, found elsewhere only once in Prudentius,55 and in No. XCIV, occurs a phrase somewhat resembling one used by that same author. But these instances are too slight on which to postulate a direct borrowing from the Christian poet (with whom, however, Symphosius may well have been contemporary).

 p20  The enigmas of Symphosius are thoroughly pagan in character; the subjects are drawn from the external world and include for the most part objects closely associated with the daily life of man. They fall in no definite groups, though there is a tendency to associate those dealing with similar or related subjects. Many deal with animals; others with plants, flowers, vegetables and food-stuffs; many more with articles of clothing and personal adornment, tools and implements of domestic use, structures of everyday use, such as ship, bridge, ladder, etc. and finally, several with natural Philadelphia such as clouds, rain, snow, etc. The range is considerable and the variety interesting. Some are easily guessed; others (such as, e.g., No. XCIV, One-Eyed Garlic Vendor) would be impossible of solution without the lemmata. How many of the riddles are strictly original with Symphosius we have no way of telling, for it would be quite wrong to infer that all those other than the half-dozen or so for which parallels can be found in Greek literature56 originated with him.

§4 The importance of Symphosius; his successors

Although we cannot accurately gauge our indebtedness to Symphosius as the originator of his enigmas, yet it is to him that we owe their preservation in a poetic form of no mean merit. He has been rightly termed "in one sense the father of the riddles of our era."57 He is to riddle-writing what Martial was to the epigram; he gave its artistic form and set the standard for future generations. As Tupper says, "The enigmas of Symphosius have dominated all riddles, both artistic and popular, since his day."58 He set the fashion for writing them in groups of 100, a fashion which still persists in such a modern title as, e.g. "A Century of Charades."59 His enigmas influenced widely  p21 all riddle-writing of the Middle Ages. We find the first traces of this influence in the anonymous Historia Apollonii Regis Tyri,60 a story which enjoyed a tremendous vogue throughout the Middle Ages. In the best recensions of this tale61 no fewer than ten of the enigmas of Symphosius are to be found.62 At least three these passed with the Apollonius story into the Gesta Romanorum.63 In the Disputatio regalis iuvenis Pippini cum Albino scholastico we find Alcuin paraphrasing seven of Symphosius' riddles for the instruction of his royal pupil.64 In the important and interesting group of 63 hexastich Latin riddles found anonymously in several MSS. (the oldest and best of which is Bernensis 611 s. VIII‑IX, hence they are usually referred to as the "Bern Riddles"; for text see Riese, Anth. Lat. No. 481) ten deal with the same themes as Symphosius, but in a totally different and original fashion. Yet even here occur phrases that show the influence of the earlier writer.65 And in the great Anglo-Saxon riddle collection, the Exeter Book, we find direct imitation of Symphosius both in subject-matter and style.66

But it was among the Latin writers of riddles in England in the eighth century that the influence of Symphosius reached its height. Of these the greatest by far was Aldhelm of Malmesbury (640‑709), Bishop of Sherburne, who, following the fashion set by  p22 Symphosius, wrote a century of riddles in hexameters. Unlike the enigmas of Symphosius, they are of varying length. In them there appears, of course, a Christian element totally lacking in Symphosius. Yet there are many references to classical mythology, and many of the subjects are those of Symphosius. Several enigmas are directly borrowed, as well as many a striking motive or phrase.67 But Aldhelm's chief indebtedness is to be found not so much in his enigmas as in the Epistola ad Acircium or Liber de Septenario, a treatise on prosody which serves as a prose preface to the riddles. In this work, which he sent in 1695 to Ealdferth, King of Deira and Bernicia, he acknowledges his indebtedness to Aristotle and to the Old Testament, but chiefly to Symphosius, for whom he expresses the highest regard. He gives us our first direct reference to Symphosius by name and praises him in the following terms:68 Nam Symphosius, poeta versificus, metricae artis peritia praeditus, occultas aenigmatum propositiones, exili materia sumpta, ludibundis apicibus legitur cecinisse, et singulas quasque propositionum formulas, tribus versibus terminasse. In the course of his rather lengthy treatise he quote Symphosius no less than a dozen times, twice by name and ten times without indication of source.69

The next great name among English ecclesiastical riddlers of the eighth century is that of Tatwine, Archbishop of Canterbury († 734), who wrote 40 riddles in Latin hexameters, but like those of Aldhelm, they were of varying length. He owes distinctly less than Aldhelm to Symphosius, as the influence of the latter can be traced in only about six instances, and these may be due more to similarity of subject than to actual borrowing.70 But several phrases he has taken directly from Aldhelm. Acquainted  p23 as he undoubtedly was with the enigmas of Aldhelm, he can scarcely have been unaware of Symphosius, whom Aldhelm hails as his great exemplar in the art of riddle-writing.

Shortly afterwards, the forty enigmas of Tatwine were supplemented by Eusebius (probably, though not certainly, the great Hwaetbert, Abbot of Wearmouth in Northumbria), who wrote a complementary sixty to round out the hundred, thus continuing the tradition established by Symphosius and followed by Aldhelm. In four instances there is some evidence of the influence of Symphosius; in sixteen that of Aldhelm.71

At some time in this period, in all probability, are to be placed the Flores of the Pseudo-Bede,72 in which five riddles from Symphosius are quoted in full, though in some garbled form.

In the twelfth century Sigebertus Gemblacensis, in his de Scriptoribus Ecclesiasticis c. 132, mentions Aldhelm, in the following terms:73 Adelmus episcopus, imitatus Symphrosium, qui per prosopopoeiam qualitates singularum rerum exprimens scripsit librum aenigmatum metrice, exprimens et ipse qualitates rerum, scripsit aenigmatum librum, et in mille versibus consumavit illum. In addition to this testimonium we have that of the author Anonymus Mellicensis, who in his de Scriptoribus Ecclesiasticis c. 731 mentions Symphosius with the same misspelling of the name:74 Symphrosius, vir eruditissimus sermone ac scientia sua, nihilominus scribit aenigmata, quae cum hodieque habeantur a pluribus, nos tamen necdum eorum legendorum copiam habere potuimus.

The enigmas attracted some attention during the Renaissance. About 1540 they were translated into Greek by Joachim Camerarius, who published 17 of them in his Elementis Rhetoricae.75 In 1602 they were published with expansions and additions by Reusner's pedants.76

 p24  §5 Editions and Translations

The editio princeps was by Joachimus Perionius, Paris, 1533, who included also the anonymous Cuculus riddle, which he dropped from his 2nd ed. in 1537. In a brief and quaintly worded preface, after expressing his delight in chancing upon the work of so entertaining an author and expressing his desire to save this excellent work from oblivion, he passe judgment upon Symphosius in terms which still hold good: De puritate quidem Latini sermonis, cum iis qui supra mille annos scripserunt . . . facile contenderit. Paucis vero antiquorum palmam apte dicendi concesserit, ut qui res obscurissimas tractet planissime, in quo vis ingenii perspici potest vel maxima. Iam vero rerum earum quas tractat, quod artem magna ex parte in iocos, risus et sales contulit, in quibus peritia sola dominatur. Next came the Roman editions of Castalio (1581, 1597, 1607). The enigmas were included in the Poemata Vetera of Pithoeus, Paris, 1590, a work which was often republished (e.g. London 1596, Geneva 1619), and this remained the standard text until Heumann was led by his theory to publish them under the name of Lactantius in 1722. But by far the best editing they received was by Wernsdorf in 1799 in his Poetae Latini Minores. His text was printed by Lemaire in the Bibliotheca Classica Latina, Paris, 1826, with a Latin commentary of admirable thoroughness. This was the best text available until that of Riese in his Anthologia Latina in the Teubner series (1869; 2nd ed. 1894). The text occurs a second time in the Teubner series in the Poetae Latini Minores of Baehrens (1879; 2nd ed. 1892). Of the two, Riese's text is by far the better; his apparatus criticus is fuller and more exact, and his text does not suffer from the excessive and capricious emendation that mars the work of Baehrens.

The enigmas as a whole have been translated but twice: once with a charming display of Gallic wit and a deft touch by Prof. E. F. Corpet, Paris, 1868, and again by Prof. Elizabeth H. du Bois (Peck) in Vermont in 1912. Her English rendering (like the French version of Corpet) is in verse, the exigencies of which at times preclude strict accuracy in translation. Her fidelity to the text of Baehrens, emendations and all, as well as several  p25 mistranslations and an inadequate commentary, detract from the work as a whole, even as an attempt to 'popularize' the author. It is in view of the lack of an exact prose rendering in English77 and of an adequate commentary (other than the century-old Latin one of Lemaire's ed.) and of an entirely satisfactory text (by either Reise or Baehrens) that I have attempted this edition of text, translation, and commentary.

§6 Text and manuscripts

The complete text of Symphosius is preserved in 16 MSS. In addition there are 4 MSS. which preserve the text in part, of which only one, s (Parisinus 4808, ninth century), has any value. By far the most important of the extant MSS. is A (Parisinus 10318), the great Salmasian Codes, so called from its former owner Claude de Saumaise (Salmasius), into whose possession it came between 1609 and 1620. It is written in uncials of the late seventh or early eighth century. It contains corrections in the original hand (A1), as well as many marginal emendations by Salmasius himself, who was a thorough scholar of penetrating insight and sound judgment.78 The remaining MSS. are divided into two families, B and D. The most important MSS. of the B class are β and w, both of the tenth century; despite the age of M (eighth century), the most important MSS. of the D class are d (ninth or tenth century), α (ninth century) and h (tenth century). In my apparatus criticus, which follows in general the lines laid down by Riese, I have given fully, from his edition, the readings of Aβαd, and of the other MSS. such readings as were pertinent. With this has been compared the apparatus criticus of Baehrens, whose readings of the chief MSS. agree in every single instance with those of Riese. I have included all the emendations of Baehrens and Froehner, however improbable, together with such emendations of early editors from Perionius to Wernsdorf as seemed of interest. Wherever such an emendation has been  p26 incorporated into the text it has been noted in the commentary. I have also included in my apparatus criticus all the readings of Riese which I have failed to incorporate in my text.

Heumann and Wernsdorf simply listed the MSS., chiefly in order of age, making little attempt to distinguish between families. It was Riese what first, in 1868, in a brief by admirably thorough article,79 placed the textual criticism of Symphosius on the sound basis of modern scholarship. He noted the existence of two separate recensions. Of these he believed B to be the older, a 'first draft' from the author's own hand, and D a later revision by Symphosius himself. This opinion he reiterates in his first edition of the Anthologia Latina the next year, but in his second edition of 1894 he added a significant paragraph in which occur the following words:80 Recensiones D et B ab ipso poeta profectas olim dixi; quod nunc non tam fidenter profero. For it may be said in general of the readings of B that they are often less felicitous than those of D, in a few instances they are quite unmetrical; they seem, too, to have suffered greater corruption in transmission. On the other hand, while none of the versions of D are unmetrical, many are at the best but obvious attempts to better something not quite fully understood in B. In several instances rather pedantic phrases or even whole lines have been substituted for the earlier recension, and in one case an entire enigma.81 These might better be reckoned to the credit of later scribes and emendators than to Symphosius himself.

If A, our oldest and best MS., were definitely aligned with one or the other recension, the weight of its authority would settle most of the textual problems. Unfortunately the scribe of A had at his elbow both recensions, for in many places the reading of B is immediately followed by that of D with no break in continuity.82 The scribe of A was literal and unimaginative;  p27 he made no attempt to better his text or even to have it make sense, but faithfully reproduced the versions of B and D, errors and all. But, be it noted, in all instances where both versions are given that of B precedes, and in the far more frequent instances where only one version is given the reading is almost (not quite) invariably that of B, not D. This combined weight of AB is the deciding factor in Baehrens' text, but Riese believed so strongly in his theory that D went according to Symphosius himself that he has permitted the readings of D to displace those of B in every instance, even when the result is far from happy. This accounts, in part, for the wide difference between the texts of Baehrens and Riese. But I do not believe that a satisfactory text can be arrived at through either a rigid adherence to the readings of D or a decided preference for the readings of B, marred by liberal and capricious emendation. B would seem to be slightly older than D, yet it fails to preserve the true reading in many instances. On the other hand, B not infrequently preserves what must have been the original reading as against an obvious substitution. In many cases the conflicting readings are of so nearly equal value that a decision is difficult. Here I have adopted the eclectic method of earlier editors from Perionius to Wernsdorf, who with all their faults, often produced a truer text than either Riese or Baehrens, since they were not hampered by any theories as to the relative value of the MSS. They followed in the main AB, but did not hesitate to use the readings of D when they were demonstrably better.

Riese in his edition of 1894 made no alterations in his text or apparatus criticus. Yet in his introduction he recedes from his earlier position that D represents an 'improved' version turned out by Symphosius himself, and makes a modified and more cautious statement, to which I can subscribe: Equidem si non ab ipso poeta, at certe ab aequalibus eius vel brevi post, cum eius aenigmata etiamtum maxime vigerent, verba ludendi causa (idque interdum et praecipue in B non satis feliciter) immutata recensionesque quae in A contaminatae, in DB separatae sunt ortas esse existimo. Had he reflected in his text his altered readings, he would undoubtedly have abandoned some of his readings.

 p28  In constituting my text I have endeavored to steer as sound and conservative a course as possible, rejecting all unnecessary emendations, whether of Baehrens or Riese or earlier editors. The very nature of Symphosius' subject makes it necessary that the meaning of many lines is enigmatical and not immediately apparent. But comparatively few lines are really obscure. Much ingenuity has been spent by certain scholars in emending the text of Symphosius, where others have bent their efforts toward gaining the meaning of the verses as received, if at all possible. And except for two or three places where the text is obviously corrupt in all the MSS., I feel that the readability of even the most difficult lines can be established.

 p29  SIGLA83

A = olim Salmasianus, nunc Parisinus 10318 (= Baehr. S)
(olim supplementi Latini 685)
s. VII vel VIII init.
Salm. = emendationes Salmasii s. XVII (= Riese a)
s = Parisinus 4808, frag. vv. 1‑74 s. IX
B = recensionem quam exhibent (= Baehr. β) s. X
β = Sangallensis 196 s. X
K = Parisinus 2773 s. XI
L = Parisinus 8088 s. XI
w = Westmonasteriensis E 919 s. X (= Heumann B)
D = recensionem quam exhibent (= Baehr. β)
d = Vossianus quart. 106 s. IX‑X
α = Sangallensis 273 s. IX
F = Parisinus 5596 s. IX
G = Parisinus 5596 s. X
H = Parisinus 8440 s. X
I = Parisinus 8319 s. XI
M = Petropolitanus F XIV, 1 s. VIII
N = Palatinus 1719 s. IX
O = Palatinus 1753 s. IX
h = mus. Britann. Regius 12 C 23 s. X (= Heumann A)
Ang. = Angelicanus V 3, 22 s. XI

Numera suprascripta (e.g. A1, α1, β1,) indicant manum primam vel secundam.

The Editor's Notes:

1 Cf. F. B. Gummere, Beginnings of Poetry, N.Y., 1901, pp451‑452: "Metaphors of the substantive may well have been the origin of the riddle."

2 Amer. Jour. Psych. VIII (1897‑1897), 484.

3 Rhetoric III.11.6 (quoted from Welldon's transl., Lond., 1886, p264).

4 Judges XIV.5‑18 (quoted from the Holy Scriptures, Jewish Pub. Soc., Phila., 1917); Flav. Josephus Antiq. V.8.6.

5 First Kings, X.1‑3.

6 Cf. Friedreich, Geschichte des Rätsels, Dresden, 1860, pp98‑103.

7 Fl. Joseph. Antiq. VIII.5.3.

8 A good text is found in the Sammlung Mittellateinischer Texte vol. 8, Carl Winter, Heidelberg, 1914.

9 For an English version cf. Kemble, Salomon and Saturn, 1848.

10 Cf. Wilmanns, Haupts Zeitschr. f. Deutsch. Altertum XIV.530.

11 Ibid.

12 cf. Ohlert, Rätsel und Rätselspiele der alten Griechen, Berlin, 1912, pp17‑22, with many citations and references. This excellent work (the outgrowth of an earlier admirably thorough monograph, Rätsel und Gesellschaftspiele der alten Griechen, Berlin, 1886) is a veritable mine of information and one of the most important contributions to riddle-literature to date.

13 cf. Ohlert, pp30‑32 and 34‑46.

14 cf. Buechler, Das älteste Latein. Rätsel, Rh M 46 (1891), pp159‑160; also the excellent English rendering of Rolfe in the Loeb ed. of the Attic Nights of Aulus Gellius (vol. II, N.Y., 1927).

15 For similarity of ideas involved in the play on the word caput cf. Symph. XCIV.2.

16 Symph. XCII.2 closely parallels this proposition, esp. in his use of the word anima.

17 cf. Ohlert, pp98‑99.

18 cf. Friedreich, pp98‑99.

19 cf. note to Symph. Praef. 3.

20 Cf. note to aenig. XC.

21 Yet cf. aenig. XXXVI and LXXIV.

22 De Viris Illustribus, c80.

23 cf. note on Symph. Praef. 1‑2; Goetz, Rh M 41 (1886), pp318‑319.

24 Poetae Latini Minores, Helmstadt, 1799, VI, 424.

25 Dissertatio de Symposii Aenigmatis (Part I), Berlin, 1854, p14. Beginning with the edition of Castalio (Rome, 1581) the variant spelling Symposium, which is found only in 2 MSS., is accepted by several editors and commentators (e.g. Lemaire, Bibl. Class. Lat., Paris, 1826).

26 In Smith's Dict. of Gk. and Rom. Biog. and Mythol., London, 1849, we find Symphosius wrongly listed under the name "Firmianus".

27 Poematia Vetera, Paris, 1590.

28 cf. note on Symph. Praef. 1‑2; Riese, Zeitschr. f. Oesterr. Gymn., 1868, p483, n1.

29 cf. Teuffel, Hist. of Rom. Lit., 1892, §449, 1.

30 Sitzungsber. der Wien. Akad. 43 (1863), p12.

31 p414.

32 p36.

33 p12.

34 De Re Metrica, p55.

35 Zeits. f. Oes. G. 1868, pp484‑485.

36 Riese, Anthologia Latina, ed. Teubner, Leipzig, 1894, Praefatio, pp. XXIV‑XXV; Schubert, Quaestiones de Anth. Cod. Salm., Pars I, De Luxorio, pp17 ff.

37 Antike und Mittelalterliche Rätselpoesie, Biel, 1869, p23.

38 Anth. Lat., Praef., pp. XXVI‑XXVIII.

39 Teuffel, l.c.

40 Dill, Roman Society in Last Century of Western Empire, London, 1906, pp385‑387.

41 Teuffel, l.c.; Schanz, Geschichte der röm. Litt., Muellers Handbuch der Altertums-Wissenschaft, Munich, 1920, VIII, 4, 2, §1036.

42 cf. Index Librorum, ed. Teubner, Leipzig, 1904, p99.

43 cf. commentary on praef. 1‑2.

44 cf. the views of L. Mueller, l.c.

45 Praef. 16; LIII.2; LXIX.1; LXXIII.3; LXXXV.3; XCII.3. There are no examples of hiatus in the text as herein constituted, though two examples occur among the readings of B, viz. LXXVII.2 and LXXXIII.3.

46 LVIII.2.

47 XVI.3.

48 LXXXVIII.1. Several of the readings of B, however, are unmetrical, e.g. LXXV.3; LXXXV.3; LXXXVIII.1; XCIII.2.

49 Yet to the compiler of the article on "Firmianus" in Smith's Dict. Biog. (l.c.) the riddles are "insipid".

50 Rh M 48 (1893), pp474‑475.

51 e.g. XXVI.1; XXXIV.1; XL.2; XLII.1; XLIII.1; L.2; LV.3; LIX.1; LXI.3; LXIX.2; LXXI.2; LXXV.3; XC.3; XCIV.1 and 3; XCVIII.3.

52 Other striking examples of this phrasing may be noted in XIV.1; XIX.1; XX.1; XXI.1; XXV.1; XXIX.2; XXXV.1; LII.2; LXIII.2; LXXXVIII.1.

53 The answer is: Clepsydra.

54 The nunc is first word of the third line only in AD. For metrical reasons I have preferred for my text the reading of Bh.

55 There is one other instance in Carm. Epig. 1327.

56 E.g. those on Smoke (VII), Vine (LIII), Ball (LIX), Saw (LX), and Sleep (XCVIII) repeat the queries, if not the form, of several enigmas in the fourteenth book of the Palatine Anthology. But such queries, like the riddle on the Louse (XXX), were common currency.

57 Tupper, The Riddles of the Exeter Book, Boston, 1910, Introd. p. XVII.

58 l.c., p. XXX.

59 Wm. Bellamy, Boston, 1896.

60 The best ed. is that of Riese (Teubner), Leipzig, 1893.

61 As we have it, it is probably not older than the 7th cent., though the original of this tale is to be sought in Hellenistic literature. The enigmas represent, then, the embellishments of the medieval scribes who refashioned the story.

62 In chapters 42 and 43 the following enigmas occur: XII, II, XIII, LXXXIX, LXI, LXIII, LIX, LXIX, LXXVII, LXXVIII (in order of occurrence). Each instance I have noted in my app. crit. to the text of Symphosius.

63 In chap. 153 occur Nos. XII, LXXXIX, and XIII.

64 cf. Wilmanns, Haupts Zeits. f. Deut. Altertum XIV, 530. The enigmas used are Nos. LXXV, XXX, XIV, XCVIII, XCIX, XII, XCVI.

65 E.g. Symphosius says of the Cepa (XLIV.1): mordeo mordentes; Bern, De Pipere (XXXVII.5): mordeo mordentem. There are few examples of direct borrowing, but dozens of lines clearly 'echo' the phrasing of Symphosius.

66 cf. Tupper's ed. sup. cit. and Wyatt, Old English Riddles, Boston, 1912. The following enigmas are either directly imitated, or have influenced in greater or less degree the riddles of the Exeter Book: LVI, LXI, LXXII, XVI, II, XLIV, XII, XCII, IV, as well as the enigma on the Cuculus (falsely ascribed to Symphosius; v. my note at end of text). The notes of both Tupper and Wyatt contain full discussion of the extent of the debt to Symphosius in each instance. Wyatt translates in his notes the riddles enumerated above.

67 Paul, p19, has a list of the most obvious parallel passages; cf. Manitius, Zu Aldhelm und Baeda, Wien, 1886, p51 and pp78 ff., where the extent of Aldhelm's indebtedness to Symphosius receives its fullest treatment.

68 Quoted from Migne, Patr. Lat., vol. 89, 170C.

69 The passages quoted are XLVIII.1; XCVIII.2; LXIII.1; XVII.2; LXXXIV.3; XCI.3; LVIII.3 (twice); XXII.3; LIII.3; XXIV.3; XXXVI.2; LII.2.

70 For a complete discussion of Symphosius' influence (actual and fancied) upon Tatwine and Eusebius, as well as upon Aldhelm and the Bern Riddles, cf. Ebert, Die Rätselpoesie der Angelsachsen, Ber. über die Verhand. der sächs. Gesell. der Wiss. zu Leip., Phil.-Hist. Classe, 1877, pp20‑h56.

71 cf. Ebert, l.c.

72 Text in Migne, vol. 94, 543‑548.

73 Text in Migne, vol. 160, 576A.

74 I have been unable to find this author in the indices to Migne. But the passage quoted is given by Paul, and occurs in Lemaire's ed. of Wernsdorf's text of the Poet. Lat. Minute. vol. VII, p341.

75 Basel, 1541, p313.

76 Reusner, Aenigmatographia sive Sylloge Aenigmatum et Griphorum Convivialium, Frankfurt, 1602.

77 Ten of the enigmas have received prose renderings by Wyatt in his notes to his Old English Riddles, pp66‑123.

78 For full description with account of its history cf. Riese Anth. Lat. praef. pp. XII‑XXXI; also Traube, Philologus vol. 54, p124. It has been completely reproduced in facsimile by the Bibliothèque Nationale, Paris, 1903.

79 Ueber die Textkritik des Symphosius, Ztschr. f. Oest. Gymn. 1868, pp483‑500.

80 praef. p. XLVI.

81 These instances are fully discussed in the commentary at the points of their occurrence.

82 For a full discussion of this, with citations of passages involved, cf. Riese's article pp485 ff.

83 Except Salm., the sigla are those used by Riese.

Page updated: 25 Aug 12