Marcus Terentius Varro’s De lingua Latina is a systematic inquiry into the Latin language by one of Rome’s most learned scholars of the first century BCE. Written in twenty-five books and dedicated in part to Cicero, it treats etymology, the formation of words, and the great Greek-derived controversy over whether language follows regular patterns (analogy) or is governed by irregular usage (anomaly). The work survives only in part: books five through ten are extant, some of them lacunose, while the rest are known through fragments and later citations. This edition, revised by Georg Goetz and Friedrich Schoell for Teubner, presents the surviving text with an extensive critical apparatus, fragments of Varro’s other grammatical writings, and a substantial editorial introduction.
“Casu fere factum est, ut societatera ad Plautum recensendnm olim a RiTSCHELio praeceptore dilectissimo iunctam ad Varronis libros de lingua latina transferremus.”
Varro approaches language as a craftsman approaches his material. Words, for him, are not arbitrary counters but signs whose origins can be traced, whose forms can be classified, and whose behavior can be reduced to rules—or shown to resist them. The treatise moves from the concrete to the abstract: from the derivation of individual words, through the mechanics of declension and conjugation, to the philosophical question of whether any system governs speech at all.
How Goetz and Schoell Built the Edition
The volume opens with the editors’ own account of how the work came to be. Goetz and Schoell explain that chance brought them to Varro after their earlier collaboration with Ritschelius on Plautus: one had offered the publisher a Teubner edition, only to find the other had already agreed to the same task. They resolved to work together, though their disputes were often sharp until one gave way, a fresh view emerged, or the matter was left undecided with a frank non liquet. They deliberately refuse to mark which editor contributed what, so that no claim outruns the evidence and nothing stands as asserted without concession.
What began as a revision grew into something larger. A fresh collation of the Laurentian codex proved necessary, and so did the gathering of testimonia—not only the relatively few explicit citations of Varro but parallel passages and ἀντίγραφα as well. They excerpted the more important witnesses and merely indicated the rest, admitting that more could have been added, since they began collecting late. Fuller prolegomena, notes, and expanded indexes followed. They kept their plan to append fragments of Varro’s other grammatical writings, even though Funaioli had meanwhile included these in his Grammaticae Romanae Fragmentorum (1907) and Wessner was preparing a collection of all Varronian remains; they explain their use of Wilmanns’ numbers in parentheses and Funaioloi’s in brackets, with a conspectus on page 242.
Wissowa and Skutsch had hoped to proofread, and some of their comments appear under the sigla Wi and Sk, but both withdrew when they took on the offices of rector and dean as the sheets left the press. Wessner could not give the help he intended because of a change of position; Adolf Groth, learning from Teubner’s announcements that the editors were publishing these books, shared old notes, of which only a few are mentioned at his request. The editors thank these men along with Diehl and Vitelli, and beg pardon that despite much time spent they could not avoid errors, noted as far as possible in the Addenda and Corrigenda. The preface is signed at Jena and Heidelberg, November 1909.
Did Varro Publish the Books Himself?
The prolegomena take up a scholarly quarrel: against Mueller, Lachmann, and Ritschelius, the editors argue that Varro did not leave the twenty-five books unedited. Following Spengel and Wilmanns, they find no reason to deny that Varro himself published them, since Jerome’s index clearly attests that he made an epitome—unlikely before the books were finished. To those who find it surprising that syntax took twelve books, Wilmanns answered that Chrysippus left far more on the same subject. Spengel plausibly conjectured that Varro first issued the triad addressed to Septumius (books II–IV), since otherwise he would have dedicated the whole twenty-five books to Cicero rather than twenty-two; the first book was prefixed to the completed work, not the triad, and the opening of book five supports this. Vahlen discussed Varro’s manner of addressing Cicero, whom he names only in individual books’ titles. Varro also fashioned an example in Cicero’s honor, “consul fuit Tullius et Antonius,” just as he sometimes puts “Marcus” and more often “Terentius” as an example for his own name, and uses “Reatinus,” often referring to the Sabines.
Of the books’ early fortunes after publication, little or nothing is known. Passages in Livy, Vergil, Propertius, and Ovid that seem to go back to Varro recall the Antiquitates rather than the grammatical works. Vitruvius knew the books before 27 B.C., and his words are prefixed to the first page. Mueller showed that Verrius Flaccus, who often agrees with Varro, seems to have drawn his learning elsewhere—evidently from the same books, especially the Antiquitates, from which Varro himself drew in grammatical matters; Verrius’ methods were quite different from Varro’s. Remmius Palaemon, pioneer of a new way of treating Latin grammar, may well have read Varro, though certain traces are lacking in his fragments; whether Suetonius’ report that Palaemon called Varro a pig—an epithet chiefly marking a rough style of speaking, per Norden’s probable conjecture—refers to these books is uncertain. Perhaps what Quintilian says, not without mockery, about Varro’s etymologies in book five goes back to Palaemon as teacher, though Spengel supposed Quintilian read Varro himself.
Stilo’s Shadow and the Second Triad
The editor argues that testimonies from Plutarch and Dionysius, who depend on the same passages of the Antiquitates, together with weighty testimonies from Cicero agreeing with Varro, allow firmer judgments. He finds it remarkable that Spengel, at V 68, adduced similar words from Firmicus Maternus without noticing that Cicero’s De natura deorum was their source. The consensus is best explained by positing that the Antiquitates were used—yet Varro did not always transcribe verbatim; in excerpting he accommodated words to grammatical method and arranged details as convenience served. Nor did he draw only from the Antiquitates but inspected other works of his own; even in the seventh book such passages, though rarer, are not wholly absent. He then attacks Reitzenstein’s view that most material is not Varro’s but Aelius Stilo’s, which Varro made his own and often corrupted: for however many passages derive from Aelius, in just as many other matters Varro plainly does not depend on him, and the foundations of Reitzenstein’s edifice collapse.
The second triad, books VIII–X, stands in far different condition. Beginning with book VIII, it concerns analogy and anomaly, a controversy received from the Greeks that stirred keen studies among the Romans. In book VIII, which argues against analogy, the first part (1–24) differs in more than one respect from the rest; others and Roehrscheidt saw it as proceeding from an analogist rather than an opponent, and this introduction belongs not to book VIII but to the whole dispute. Varro began what he wished demonstrated at paragraph 25, teaching why and in what words are declined, then the two kinds of declination, voluntary and natural. He passes to objections against universal analogy (25–43), drawn from utility, elegance, custom, the contrary, the fact that analogy is not everywhere, and the poor answer defenders give about similarity; then against individual parts of speech—provocabula, vocabula, nomina—with the rest lost to a lacuna except fragments Gellius preserved on verbs, conjunctions, and adverbs.
Disertus, Sermo, and the Hand Laid on the Matter
The edition opens with a dense layer of testimonia and critical notes, citing Festus, Isidore, and the Servius Danielis commentary to pin down the meanings of disertus and sermo. Festus derives disertus from disserendo; Isidore from dissertim disserit; Servius explains edissere as narrating all things in order, whence plain speakers are called dissertores. Sermo itself, Servius says, is a joining-together of speech between two or more, properly what is “sown between” both parties. The notes also gather Cicero and Gellius on manum consertum, the old legal phrase by which litigants summoned each other from law to join hands over the matter at issue—Ennius’s line “non ex iure manum consertum sed magis ferro” is cited—and Festus on asserere manum, the gesture of laying a hand on someone to lead him from slavery to freedom.
What Legere Gathered
Varro then works through the family of legere: legitima, collegae (who are una lecti), sublecti, allecti, collecta, and ligna, because fallen wood was gathered for the hearth; thence legio, diligens, dilectus. He turns to words named from similarity of sound—murmuratur, said of one who speaks so lightly that he seems to do it from sound rather than to be understood, whence the poets’ “murmurantia litora”; and likewise fremere, gemere, clamare, crepare. Nearby stand quiritare, the cry of one imploring the faith of the Quirites (the name coming from the Curenses, who entered partnership of citizenship with the Romans under King Tatius), and iubilare, its rustic counterpart, illustrated by a comic line: “o bucco! quis me iubilat? — vicinus tuus antiquus.” Triumphare is named from the soldiers who shout “io triumphe” through the city as the commander goes to the Capitol, perhaps from the Greek θριαμβος, surname of Liber.
Plautus and the Poets as Witnesses
Book VII proceeds by quoting the early poets for each word. From the Persa, congerro derives from gerra, a Greek word. From the Menaechmi, ascriptivi are so called because unarmed men were formerly enrolled to supplement armed soldiers who had died. From the Trinummus, ferentarius comes from ferendo, or names cavalry with only throwing weapons—Varro says he saw such horsemen painted and labeled in the old temple of Aesculapius. From the Frivolaria, rorarii are named from dew, because they began battle before rain; accensi are attendants, per Cato. Dividia means a distraction of grief; syncerastum is an old Greek word for a dish; trames is from transverso; limax from the mud where it lives; diobolares means worth two obols; schoenicolae from a cheap perfume; miraculae from miris, whence Accius’s miriones. Scrattae come from excreando; scruppeda is variously explained as scauripeda or a hairy worm or “rock-foot,” whence Accius’s scrupea for difficulty. Axitiosae, per Claudius, are fellow-pleaders from agendo, as factiosae from faciendo; scobina is from sawdust; gralator from gradu magno.
The Mourning Woman and the One-Eyed
A praefica is a woman hired to praise the dead before the house—Aristotle records the custom, Claudius says she directed the maids’ lamentation—and both show the name comes from praefectione. Ennius’s Coclites yields cocles from oculus, the one-eyed, as in the Curculio.
The Measure of the Night
Time words follow. Intempesta nox comes from tempestas, from tempore—the time when nothing is done. Ennius’s temo measuring the night prompts Varro’s reasoning that rustics first marked sky signs for agriculture; the Greeks call the seven stars the wagon, the Latins triones (from oxen, or terriones from earth), temonem (from tenendo), and axem. Iubar is the morning star, diffusing light like a lion’s mane, whose rising signals the night’s end. Crepusculum is a Sabine word for the doubtful hour; concubium from concubitu; conticinium from falling silent.
A Final Sweep of Words
The book closes with a rapid run: reciproca from recipere; transversus against proversus, with Plautus’s crab image; Ennius on Andromache and Paris, imitating Euripides but lapsing; aurora from the sun’s golden air, and rutilare from the same color; scortari from leather goods; numen from nutu; epityrum, a Sicilian food; lymphata from lympha; the alcyon days of calm sea; rite as recte; comiter from Greek komos; cicurare, to tame; ferme from ferendo. Interjections like euax and heu signify nothing but natural effusion; iurgare is from iure; clepere from clam or Greek kleptein; mandier from mandendo; obscenum from scaena, since scaeva, the left, once held good omens; crevi means constitui of an heir; frequens is assiduus.
The Case Against Analogy
Book VIII turns to the argument that similarity should be preserved in forming words from names: since Parma, Alba, and Roma are alike, we say Parmenses; from Roma, Nola, Parma come Romani, Nolani, Parmaui.
Three Kinds of Style and the Athenian Embassy
The fragments gathered under the heading De sermone Latino open with a theory of style in three “probable” kinds answering to the Greek terms: the first called uber, the second gracile, the third mediocre. Amplitude is the dignity of the copious kind; charm and subtlety belong to the slender; the middle shares in both. Each virtue drags a counterfeit twin behind it—swollen and inflated things pass for copious, squalid and jejune things for slender, uncertain and ambiguous things for moderate. Varro is said to have set the true examples in Latin: Pacuvius for abundance, Lucilius for slenderness, Terence for the middle course. The same triad is traced back to Homer in three persons—the magnificent and copious in Ulysses, the subtle and restrained in Menelaus, the mixed and moderate in Nestor—and again to three philosophers.
An anecdote makes the point concrete. The Athenians sent envoys to the Roman senate to beg remission of a fine of nearly five hundred talents imposed for the devastation of Oropus. The ambassadors were Carneades of the Academy, Diogenes the Stoic, and Critolaus the Peripatetic. Brought into the senate, they used the senator Gaius Acilius as interpreter, but beforehand each had displayed his eloquence separately before a great crowd. Rutilius and Polybius admired all three, each in his own kind: Carneades violent and rapid, Critolaus shrewd and polished, Diogenes modest and sober.
Scattered Questions of Sound and Spelling
Further fragments pick at points of grammar and orthography. Interjections and the expression of feeling are treated, with Terence quoted. There are notes on phrases—domi suae, mutuo—and on clausulae, so called because they close the sentence, with Accius, Caecilius and Terence cited. The trochaic septenarius is called the “squared” verse, and the comic quadratus is formed by adding half-feet to the senarius; the transverse mark used between syllables comes in for comment. Latin is said to have twenty-three letters, of which Varro calls seventeen sufficient: H is mere aspiration, X double, K and Q superfluous, Y and Z Greek. Caesar’s habit of pronouncing and writing words like lacrumae and maxumus with I is recorded. Accius erred in doubling vowels to mark long syllables, and Lucilius was inconsistent in his spelling of I and E; Varro argues the question of singular versus plural spelling, touching forms like mium, commircium, Mircurius, miis. The digamma, its shape and its use before vowels, is discussed with votum and virgo; plebs and urbs are set against the spelling Pelops. Etymologies follow in a rush: delirus from lira, a furrow, like oxen turning from the straight course; Delmatia from Delminum; feriae from fesiae; arispex from ariga; intervallum from vallum; narare with a single R. The voice has three distances—altitude, thickness, length—measured by accent, aspiration and syllables. Questions about H accumulate: vehemens, reprehendit, halica, hortus, cohortes, harena from Sabine fasena, haedus from faedus, hircus from fircus, faba from haba; H’s position before consonants; cilonem and cocleam without aspiration; Graccus and ortus; pulchrum unaspirated; R not aspirated initially, as in Rhodum and rhetorem.
Accent, Rhythm, and the Definition of Verse
On accent the fragments ask what its matter is (the voice), its place (the syllable), its nature (up or down). Athenodorus held two prosodies, Dionysius Thrax three, Tyrannio four; Varro supports the middle prosody, arguing from music, from the mean between extremes, from the obscurity of beginnings, from the center of a circle. Earlier authorities are marshalled—Glaucus of Samos, Hermocrates of Iasius, Theophrastus, Athenodorus—with Glaucus’ six prosodies, Eratosthenes and Theodorus on the flexed accent, and Ammonius, Ephorus, Dionysius of Olympia, Hermocrates and Epicharmus naming the fourth. The notation of acute, grave and flexed accent is set out. Rhythm is distinguished from metre as matter from rule; cola, comma and period are defined, and the definition of verse closes the account.
The Combination of Propositions
The main text turns to judgments about the copulation of propositions, illustrated by “si ambulat movetur.” Whoever says “if he walks, he is moved” wants to prove something; if this is not conceded, he must either show that the person walks and reach the conclusion that he is moved, or show that he is not moved and conclude that he does not walk. A simple sentence like “homo iste ambulat,” once conceded, combined with “quisquis autem ambulat movetur,” yields the necessary conclusion “homo iste igitur movetur.”
The subject then divides into two primary parts: one concerning things said simply, the matter of dialectic, the other concerning things said in combination, its work. The first is called de loquendo; the second splits into three—de eloquendo, which fills out a sentence but makes no question; de proloquendo, which judges simple sentences; and de proloquiorum summa, which judges the combination itself until the conclusion.
Word, Thing, and the Origin of Speech
Definitions follow. A word is a sign of a thing intelligible to the hearer when uttered by the speaker; a thing is whatever is sensed, understood, or latent; a sign shows itself to sense and something else to the mind; speech is giving a sign with articulate voice. Four terms are distinguished: verbum, dicibile, dictio, res. Competing etymologies of verbum are reported—from verberare, striking the air or the ears; from verum, truth; from verum boando, Ennius’ “bombum pedum.” The Stoics claim every word’s origin can be explained, and the account traces origins through sound-similarity (hinnitus, balatus, clangor), then similarity, then catachresis (as piscina), then contraries (lucus, bellum, foedus), then vicinity—efficiency, effect, container, contained, part and whole.
The Case for and Against Regularity
The intellectual drama of the surviving books lies in the dispute between analogy and anomaly, a controversy Varro inherited from Greek grammarians and transplanted into Latin soil. Book eight marshals the arguments against universal analogy: utility resists uniformity, elegance prefers variety, custom overrides rule, and dissimilar things often produce similar forms. Book nine answers these objections, insisting that analogy exists in nature and in language alike—that within every kind, whether animals or words, similarity vastly outnumbers difference. Varro’s position is nuanced: he grants that usage sometimes deviates, but he treats such deviations as faults of individual speakers rather than failures of the art itself. The doctor does not abandon medicine because patients ignore it; the grammarian does not abandon analogy because the vulgar mispronounce.
Book ten, which opens with a promise to lay the foundations of the discipline, breaks off after distinguishing similarity from dissimilarity and outlining the four elements of verbal declension. What survives is enough to show that Varro intended a constructive synthesis, not merely a polemic.
Language as a System of Signs
Underlying the grammatical technicalities is a philosophical account of what words are and how they mean. Varro distinguishes four terms: the word itself (verbum), the thing signified (res), the dicible (dicibile, what is understood but not yet uttered), and the utterance (dictio). A word is a sign of a thing intelligible to the hearer when spoken by the speaker. This fourfold scheme, which the text attributes to Stoic sources, allows Varro to separate the physical sound from the mental content and the external referent—a distinction that makes possible the systematic study of meaning and derivation.
His etymologies proceed by recognizable principles: from sound-similarity (the neighing of horses, the bleating of sheep), from similarity of form, from catachresis or abuse, from contraries, and from proximity of various kinds. The results are often fanciful—lucus from lux because a grove is dark, bellum from bellus because war is not beautiful—but the method is consistent, and the ambition is clear: to show that language, however wayward, can be made intelligible.
The Work’s Place in Roman Learning
Varro’s grammatical writings circulated among the learned of Rome and left traces in later authors. Vitruvius knew them before 27 BCE; Verrius Flaccus, the great antiquarian, drew on the same sources; Quintilian mocked some of Varro’s etymologies while acknowledging his authority. The work’s influence on the Latin grammatical tradition is difficult to overstate, even where the text itself was lost. What survives—through the medieval manuscripts of the Laurentian codex, through citations in Gellius and Priscian, through the patient labor of Renaissance and modern editors—preserves a singular attempt to bring the methods of Greek philosophy to bear on the Latin tongue. Varro wrote as a Roman senator and scholar, addressing a Rome that had conquered the Greek world and was beginning to absorb its learning. His treatise is one of the earliest and most ambitious efforts to give Latin a grammar worthy of its new imperial role.
