The morphology of classical Latin

2. Concepts of morphology

If the minimal linguistic unit is really the morpheme, i.e. the minimal meaningful unit, or the minimal unit which has both contents and an expression, a signified and a signifier, as linguists say, the minimal analysis unit is actually for the linguist what we call the morphological unit, i.e. the class of the morphological segments, i.e. phonological segments which are independent from the environment. And these morphological units, alone or in combination, are used by the Latin language to form the signifier of its morphemes.

Usually, the signified of a morpheme is put in quotation marks, like “great”, the signifier between oblique strokes, like /magnus/, its phonetic realizations being in square brackets, like [maȠnus], [mag-], [maj-]. As for the morpheme, it is possible to put it in curly brackets {magnus} or to replace by its signified “great”.

The signifier of a morpheme is also called morph. A morph therefore represents a morpheme; but a morpheme can be represented by different morphs in different environments. We call allomorphs these alternative representations of a morpheme. Allomorphs are by definition combinatory variants of a morpheme, i. e. variants which are conditioned by an element of the phonological or grammatical context, and therefore are in complementary distribution. But there are also variations usually called free, which are actually variations socially or regionally conditioned.

The linguists try to classify the different allomorphs shown by a morpheme, and for that, they consider as the basis form the unmotivated form, which is then called unmarked form (or the primary form, according J. Kuryłowicz1)). So, the signified “great” corresponds in Latin to three morphs magn-, mag- and maj-; maj- appears only in context of the comparative morpheme -ior- (maior “great-er”), and mag- before the superlative morpheme -sim- (maximus “very great” or “the greatest”). Therefore magn-us is the basis form of this morpheme, since it appears in other contexts than these particular contexts. It is likely that /maj/ is not a phonological variant, but a morphological variant from /mag/ before /i/, and therefore it would be doubly motivated. As for the allomorph /mag/, it is phonetically realized like [mak] in maximus, because the phonological neutralization of voiced phonemes before an unvoiced /s/ or /t/. That can be technically formulated by the three following morphological rules:

/ mai/ / ─ [i] Comparative
great → /mag/ / ─ Superlative

If there are zero morphemes, i.e. morphemes the signifier of which is a lack of formal mark, it doesn’t mean that all lack of formal mark is a morpheme. For example, what is traditionally called present correspond to a formal lack, as we see when it is opposed to an imperfect or future:

amā-bā-s / amā- -s, amā-bi-mus / amā- -mus, amā-bā-tis / amā- -tis, amā-bi-t / ama- -t, etc.

the present tense having not the segment -bā- or ­-bi- the imperfect or future has. But does the lack of formal mark correspond to a signified? Yes, it does, as it seems: the present tense just corresponds to the present time. Possibly, but what does the present mean? It means that what the speaker says is happening. But grammars teach that the present tense can also mean the narrative past and even the future; they speak then about a historical present, which is used for the historical perfect, or a present which is used for the future, or even a conative present which denotes an action begun in present time, but never completed at all. Grammars don’t explain how the present tense can take such different and mutually exclusive values. And we don’t see what signified could be attributed to them, which would mean, according to circumstances, sometimes a present action sometimes a past action sometimes a future action. It is better to say that the present tense does not correspond to a zero morpheme; a verb in the present has not any more signified than a signifier. And if the sentence in the present is temporally situated in time, it is by a temporal adverb or, most of the time , by the type of enunciation where the sentence is found. It is actually in a narrative enunciation that the sentence on present denotes a past action, and in discursive enunciation that it can generally denote a present action, except if a temporal adverb specifies that the action is past or future. The explanation of the so-called temporal values of the present tense is not semantic, but enunciative. We certainly shall continue, for convienience’s sake, to speak about present tense, but with a small letter, in opposition to the morphological unit and therefore the morpheme of Future or Imperfect, with a capital letter.

From the same way, all the other traditional categories that have no morphological segment do not correspond to any morphological unit, as the indicative mood, the infectum aspect, the active voice, the singular number; and having no more signified than signifier, they are not any morphemes, and are spelled with a small letter.

Traditionally, linguists distinguish two kinds of morphemes: grammatical morphemes or grammemes, and lexical morphemes or lexemes.

The morphemes take the form of words, which are minimal syntagmatic units, i.e. units that can be displaced and separated in speaking; they are separated, in our writing system but not in Latin inscriptions, by blank spaces. But it is easy to identify them, a word being, in Latin, a set of morphological segments or morphological segment parts that can be together displaced in spoken string, or, as Bloomfield says, “a free form that is not constituted by minor free forms”: “a minimal free form” (Bloomfield, 11.5 p. 168).

In Latin, grammatical morphemes, like coordination or subordination conjunctions (et, aut, uel … or ut, quod, si …), and simple morphological units, like prepositions (cum, in, ab, apud …) are words; but lexical morpheme, on its own is never a word; it is a bound form. Lexical morphemes constitute words only when they are followed by at least one morphological unit, which is a case if the lexical morpheme is a noun or adjective, a personal unit, if it is a verb. So in Virgil’s verse

Maioresque cadunt altis de montibus umbrae

there are 13 morphological units (Mai-ior-es-que cad-unt alt-is de mont-ibus umbr-ae), 11 morphemes (Mai- “great”, -ior- “ more”, -es…-ae Subject and Plural ”, que “and”, cad- “fal down”,, alt- “high”, -is…de…-ibus from and Plural ”, mont- “mountain”, umbr- “shadow”), 6 words, four of which are constituted by two elements: cad-unt (V-Pers ), alt-is (Adj-AbPl), mont-ibus (N-AbPl), and umbr-ae (N-NoPl), one by one single element: de (Prep), and one by four, or rather three elements: mai-ior-es (Adj-more-NoPl) + -que (and); -que

doesn’t belong to the previous word maiores, but is prosodically linked with him; it is an enclitic (i.e. a word that always follows another word, whatever this word may be, the accentuation of which is modified, as maiō’rēsque cadunt instead of mai’ōrēs cadunt).

In Latin, only the words constituted at least by one lexical morpheme are stressed; the words constituted by one grammeme or one morphological unit are proclitics (i. e. words without accent but which lean on the stress of the following word, like de in de ‘montibus).

Some words are constituted by several morphemes, as the adjectives magn-anim-us, -a, -um “noble in spirit, brave, bold, generous”, semi-anim-us, -a, -um “half­-alive”, the nouns ama-tor, -tor-is, -tor-um “one who loves, lover,”, da-tor, -tor-is, -tor-um “a giver, donor”, impera-tor, -tor-is, -tor-um “on who gives order, a general”, moni-tor, -tor-is, -tor-um

“one who suggests or advises, a counselor”, or the verbs uēnum-dō, uen-dō “to put up for sale, to sell”, uēn-īre “to be sold”. Martinet calls this type of morphemes combination a syntheme, in order to distinguish it from the syntactical combination which is a phrase; because the syntheme functioning as a morpheme belongs to a class of morphemes but not a class of phrases, although it is constituted from two morphemes. The synthemes correspond to what Corbin calls “constructed words ”, which she sets opposite “complex no-constructed words”; the constructed words are words all the segments of which can be interpreted, which is really the case of synthemes; the complex no-constructed words are nothing but morphemes morphologically complex, but morphematically simple. It is the case of Engl. blackbird, which “looks like a phrase in two words (black bird)”, but “has to be described like a simple word (compound)” (Bloomfield, 11.6, p. 170). Why? Because “we can say black ─ I should say, buish-black ─ birds, but we cannot divide in the same way the compound word blackbirds ” (Bloomfield, 11.6 p. 171).

The term word can be ambiguous and has to be distinguished from the dictionary word. Usually, we say that the Latin word umbrae in Vergil’s verse is a form of the word umbra, it is its plural nominative, and not its singular genitive. And, of course, the dictionary gives only the word umbra. But, umbra and umbrae are pronounced and written differently, and don’t mean the same thing; they are, for the linguist, two different words, even if they are, for the dictionary, simply different forms of the word umbra. They group the same morpheme /umbr/ with two different morphological segments /ae/ and /a/, which represent two different morphological units, respectively the Nominative Plural and the Nominative.

Retour au plan ou Aller au § 3.

1) Kuryłowiz, Jerzy, 1949, “Le problème du classement des cas”, in :Biuletyn Polskiego Towarystwa Jezykoznawczego, 9, 20-26-43.