Ceci est une ancienne révision du document !

The morphology of classical Latin

1. What is morphology?

Traditionally, three fields of study are distinguished in linguistics: phonetics, morphology and syntax. In France students of Latin linguistics rely upon three well-known monographs

which indeed exemplify these three fields:

1) Phonétique historique, by Max Niedermann,

2) Morphologie historique du latin, by Alfred Ernout,

3) Syntaxe latine by Alfred Ernout & François Thomas.

Phonetics is the branch of linguistics that comprises the sounds and pronunciation of Latin. Morphology is the study of the form of Latin words. Syntax is the study of the combination of words in sentences.

The term morphology is originally a German creation (Morphologie) from the Greek morfh ‘form’.

  • Frawley, 2003, p. 82 1)Traditionally, morphology includes both the study of grammatical forms a word may take: inflection, as in Eng. form, form-s, form-ed, and word formation, the lexical relationship between words of the same family, in terms of either derivational morphology (as in form-ation, form-al), or compounding (as in cunei-form, word form)

According to this definition, one may oppose morphology, defined as the study of the internal structure of words, to syntax, understood as the study of the composition or construction of phrases, clauses and sentences. As Bloomfield writes:

  • Bloomfield p. 2072)We may say that morphology includes the constructions of words and parts of words, while syntax includes the constructions of phrases.

Following this definition,

Ernout, in the first part of his Morphologie, studies Latin declension (i.e. the forms of every case of every declension) and, in the second part, Latin conjugation (i.e. the forms of the tense affixes and the verbal stem of the infectum, which corresponds to the four conjugations of the grammatical tradition, and the formal features of the perfectum that do not correspond to these conjugations, but depend on the consonant or vocalic nature of the stem), and he ends his book with a very complete three-column index of all the studied words (p. 235-252). On the other hand in the first part of his Syntaxe, he studies the role and function in the sentence of the cases and prepositions, in the second part, the simple sentence and its constituents, and in the third part, the two means used to create a complex sentence: subordination, by which a sentence is inserted as a constituent of another sentence, and coordination, by which the constituents of a sentence can be multiplied.

A question to be considered: Modern linguists agree on the fact that the morpheme is the minimal linguistic unit of analysis, ̶ rather than the word -, as there is no universal definition of exactly what a word is. Accordingly, is the above definition of morphology still valid?

Charles Hockett’s point of view : In his Course in Modern Linguistics, Hockett defines morphemes as

  • Charles Hockett,Course in Modern Linguistics: the smallest individually meaningful elements in the utterances of a language

and proposes the following English example

  • /3ǰán+2tríjts+iz+ówldǝr+sístǝrz+vérij+2nájslij2↑/
    John treats his older sisters very nicely,

in which he finds the following list of constituent morphemes:

  • (1) John /ǰán/
  • (2) treat /tríjt/
  • (3) –s /s/
  • (4) hi- /i/
  • (5) –s /z/
  • (6) old- /ówld/
  • (7) –er /ǝr/
  • (8) sister /sístǝr/
  • (9) –s /z/
  • (10) very /vérij/
  • (11) nice /nájs/
  • (12) -ly/lij/
  • (13) /3 2 2 2/

He makes the three following remarks

  • Bloomfield, 1958, p. 126: First, intonation must not be overlooked; we have taken it as a single separate morpheme.
    Second, (5) and (9) are phonemically the same, but certainly not the same morpheme, because of the difference in meaning.
    Third, the breakdown of his into hi- /i/ and -s /z/ may seem unconvincing. The /z/ recurs, with exactly the same meaning, in John’s book, the men’s room, and the like. But the /i/ recurs only in him (as in hit‘im.)

However, there is an important point that he fails to examine: if (3) is indeed a morpheme, what is its contribution to the meaning of the utterance? What is its “signified” (Fr. signifié)? A grammarian would argue that this is a grammatical case of agreement: the subject and the verb agree in number and person. In English, with the exception of the verb be-the only English verb with different singular and plural forms both in the present and in the preterit- , the subject-verb agreement is limited to the present and the third person singular. But what exactly is “agreement”?

  • Bloomfield, 1958, p.499: Compare then: [2] verb agrees with subject a. The nurse wants to see him. b. The nurses want to see him. a. The dog was sleeping. b. The dogs were sleeping. [3] no variation in the verb a. The nurse will see you. b. The nurses will see you. a. The dog slept all day . b. The dogs slept all day. In [2] the choice between the inflectional forms wants and want, was and were is determined by the subject, whereas in [3] the modal auxiliary will and the preterit slept remain constant even though the subject changes

In this light, the segment /s/ in treats as well as in wants is not selected in order to mean a singular or non-plural person; it is determined by its subject which does not include a plural morpheme. Therefore, the form of the verb automatically depends on its context, and it is a matter of morphology, and more specifically, of morphological alternation.

Accordingly, Hockett’s example includes only the following eleven morphemes:

John treats his old -er sister -s very nice -ly /3 2 2 2 /

“John” “to.treat” “his” “old” “more” “sister” “plural” “very” “nice” “in.manner” “affirmation”

But Hockett’s conclusion remains the same:

  • Bloomfield, 1958, p.177: Morphology includes the stock of segmental morphemes, and the ways in which words are built out of them. Syntax includes the ways in which words, and suprasegmental morphemes, are arranged relative to each other in utterances.

If this definition of morphology is accurate, the definition of syntax is debatable, because it claims that words are syntactical units. In fact words, which are syntagmatic units composing the linear sequence of the spoken string, must be considered as a type of morphological unit.

Eugene Nida’s point of view : Nida3) defines morphology as

  • Nida, p.1: the study of morphemes and their arrangements in forming words,

and he writes that

  • the morpheme arrangements (…) include all combinations that form words or parts of words. Combinations of words into phrases and sentences are treated under the syntax.

He seems to have the same view of morphology as Hockett, except that he believes

  • it would be quite wrong to assume that morphology and syntax constitute airtight compartments in the structure of any language,


  • in a few instances it seems almost impossible to draw a line between word structure and phrase structure.

But what does “the study of morphemes” mean? Is it the analysis of words into morphemes, as the subtitle of Nida’s book, The Descriptive Analysis of Words suggests? Does morphology consist in identifying and classifying morphemes? This would rather be a matter of morphemics. In an article Morpheme in the International Encyclopedia of Linguistics, Peter H. Matthews says that

  • Peter H. Matthews, 2003<sup>4)))</sup>, vol. 3, p. 81: In some treatments of morphology, words are analyzed into basics units called morphemes, and the process is called morphemic.

What is morphology, then? Martinet, who defines the linguistic sign as the association between a signified, i.e. its meaning or value, and a signifier, that expresses it, and calls the minimal signs monemes, suggests a very interesting definition of morphology:

  • Martinet, 19965), p. 106: The study of the alternations of a signifier is the concern of morphology. In the traditional framework of the word or the observation of monemes, morphology doesn’t boil down to the enumeration of all the grammatical monemes in a given context: the morphology of invariable grammatical Fr. pour [Engl. prep. for ] has no relevance, whereas the lexeme all(er) [Engl. verb go], with its allomorphs /al-/, /va/, /i-/, /aj/, is a central concern of morphology.

Therefore, morphology is concerned with the forms of both lexical morphemes and grammatical morphemes, but only the forms of morphemes with alternations. With regard to nouns, Latin morphology is concerned with the five systems of casual inflexion variations, i.e. the five nominal declensions. The five declensions determine five different morphological classes of nouns. Latin morphology will focus on the alternations of the lexical morphemes that are peculiar to some morphemes of each declension.

Martinet has offered a broader definition of morphology in his course at the Ecole pratique des Hautes Etudes: he defines morphology as

  • Martinet, 1972, p. 552: the chapter of grammar which deals with all irrelevant data of the first articulation, regarding the choices of phonemes, prosodemes or respective positions.

Morphology deals with all the vicissitudes of morphemes (amalgams, discontinuities, alternations…) and gatherings of morphemes. For example, the morphology of the English or the French languages imposes the position of the determiner before the noun, while the morphology of the Romanian language imposes its position after the noun. The morphology of the German language distinguishes separable verbal particles from inseparable verbal particles. Finally, morphology is what gathers the signifiers of morphemes and constitutes word-forms, which are separated by spaces in our writing and can only be moved together. Morphology is interested in all the formal features of the first articulation that are not selected by the speaker.

Analysis of Latin into morphemes : Is the morpheme truly a minimal unit of linguistic analysis? In the first sentence of

  • Genesis, 1, 1: In principio creauit Deus caelum et terram. Terra autem erat inanis et uacua
    In the beginning, God created the heavens and the earth. The earth was without form and void.

it seems possible to find eight morphemes because of the following commutations:

  • In /a principio, in / from the beginning
  • In principio / fine, In the beginning / end
  • creauit / uidit, created / saw
  • deus / fortuna, God / Fortune
  • caelum / homines, the Heavens / the men
  • et / aut, and /or
  • terram / feminas, the Earth / the women

But what can be done with the other following commutations:

  • In principio / initio, In the beginning /star
  • creauit / deleuit,created / destroyed
  • caelum /uentum, the Heavens / the wind
  • terram / aquam, the Earth /the water.

These commutations clearly show that only the segment principi- means “beginning”, as only crea-, de­-, cael- and terr- mean respectively “create”, “God”, “Heavens” and “Earth”. But these segments never appear alone and therefore do not exist independently in the Latin language. They are always accompanied by a final affix which is - am or - a, – for the morpheme terr-, as the second sentence of this example shows:

  • Terra autem erat inanis et uacua,

but not

*Terr autem erat inanis et uacua-

or other forms like - ae, - ā, - ārum, - ās or - īs. The group of these final affixes as a whole constitutes what grammars call the first declination of the nouns. But what are these affixes? Are they morphemes? If, by signified, we understand grammatical value as well as meaning of a morpheme, we can say that the segment /-am/, which combines with the morpheme /terr-/, shows that this morpheme is the object complement of the verb created, as well as the segment /-um/, which combines with the morpheme /cael-/, and the segment /-us/, which combines with the morpheme /de-/, shows that it is subject of the verb created, as well as the segment /-a/ combined with /terr-/ in the second sentence of our example is the subject of verb erat. Therefore these segments, which have a signified, are morphemes, though they cannot commute in these examples. But there are some examples in which the commutation is perfectly possible, as in:

  • Domin-us liberauit (or occupauit)terr-am / the Lord liberated (occupied) the earth / Domin-um liberauit (or occupauit) terr-a, the earth liberated (occupied) the Lord.

But there are also many examples where the cases cannot be morphemes, but are only a part of a discontinuous morpheme, as in prepositional phrases

  • Extra urbem, out of the City

in which Kuryłowicz finds two morphemes he showed from following way:

  • extra urb- -em

The cases are therefore groups of phonemes which don’t come under the phonemic, but constitute a little morphological system of twice five markers or, more precisely, of twice five inflectional affixes.

If we examine the things more accurately, the difference is not so great between these morphological units and the signifier of morphemes, because the signifiers of morpheme can become any simple part of a signifier, as in the compound words that Danièle Corbin calls “complex words no constructed”, like fr. carpette, which doesn’t correspond to “little carp“, but means “rug”, or chaise-longue “deck chair” and chaise électrique “ electric chair”, which are not strictly speaking chairs. These compound words are simple morphemes the meaning of which is independent from the meaning of the morphological segments that constitute it.

1) International Encyclopedia of Linguistics, Oxford, vol. 3
2) Cf. Touratier, « Système des consonnes », 2005, 118-119. Alvarez Heurta, « Neutralisation consonantique en latin », 2005, 146-147, Language, London, George Allen & Unwin, 566 p.
3) E.Nida, The Descriptive Analysis of Words
4) 7
5) Cf. Touratier, 1994, « Quelques problèmes de phonologie à propos de -i- », in : Mélanges François Kerlouégan, p. 627.