Andrew Radford - Linguistics An Introduction [Second Edition]

singular

plural

nominative

feːleːs

feːleːs

accusative

feːlem

feːliːs

genitive

feːlis

feːlium

dative

feːliː

feːlibus

ablative

feːle

feːlibus

Morphology across languages

159

classes of nouns (and adjectives) is declension, and the facts of Turkish, briefly referred to above, indicate that it lacks declensions. For verbs, if we find that inflections expressing agreement, tense, etc. fall into distinct classes, as they do in Latin, we speak of conjugations (exercise 1).

Although it’s not immediately apparent, the two Latin nouns we have cited

illustrate a further characteristic feature of Latin declensions. If we look at the dative and ablative plural forms of VILLA and FELES, we find that they are

identical: viːlliːs, feːlibus. This identity obtains for all nouns in Latin, and therefore it is a fact about the grammar of Latin. Here we have to say, then, that we have a single word form but that form corresponds to two grammatical words, much

like the past tense and perfect/passive participles of English regular verbs (see p. 146). This is a widespread phenomenon in languages such as Latin, and

morphologists refer to it as syncretism. We say that the forms viːlliːs, feːlibus are syncretic, and that they syncretise the dative/ablative plural.

A rather different morphological phenomenon can be observed in Latin verbs.

In (123), we see various forms of the verb AMO ‘I love’: (123)

am-oː

‘I love’

amaːb-oː

‘I will love’

amaːb-am

‘I was loving’

amaːv-iː

‘I have loved’

amaːver-am

‘I had loved’

These forms are based on a stem form amaː- (or am- in the present tense). The final suffix is the exponent of the first person singular form, but notice that it’s a different suffix depending on the tense/aspect of the form. In the present and future forms, we have -oː but in the two past tense forms, the ending is -m, while in the present perfect form, it -iː. This kind of variation is different from that illustrated by the different noun suffixes in tables 17 and 18, because here we are dealing with forms of a single lexeme (and, moreover, an example of a completely regular verb in Latin). When we come to analyse a form such as, say, amaːviː ‘I have loved’, what we find is that the -iː suffix is not just an exponent of the property first singular – it is also telling us that the verb is present tense and perfect aspect. This is diagrammed in (124):

(124)

LOVE PERFECT PRES/FIRST SINGULAR

ama:

v i:

In (124), we can see that the property PERFECT is extended over two distinct suffixes. This situation is referred to as extended exponence (exercises 2 and 3).

An interesting fact about English is that a single base form such as walk or book is in most cases a perfectly good word. Therefore, we are tempted to think that we take the base form of a word and then add inflections to it (e.g. walk-ing, book-s),

160

words

or conversely that we can get to the base form by stripping off the inflections. This makes sense for English and a number of other languages including German,

Hungarian and Turkish, but for many inflecting languages, stripping off all the inflections often produces something which cannot function as a word. Thus, the Latin word forms we’ve seen all need some sort of inflectional ending to form a proper word. The bare root can’t be used on its own: *viːll, *feːl or *am(a) are not words in Latin. In other words, the root of a word in such languages is a bound form, not a free form. The same is true, broadly speaking, of Russian, Spanish, Greek, Latin, Japanese, Swahili, Chukchee, Navajo (for verbs at least) and many other languages. Moreover, we sometimes get a different form depending on

which set of inflections we strip off. For instance, in Latin the noun meaning

‘body’ has a basic (nominative) form corpus, but its other forms are based on the stem corpor- (e.g. corporis ‘of a body’, corporibus ‘to/from bodies’).

While English has small numbers of examples justifying more than one

stem appearing in the representation of a lexeme (see the discussion of KNIFE in

section 10), we can generally think of its inflection (or that of German, Hungarian, etc.) as being word-based, while Latin (or Spanish or Russian, etc.) inflection is stem-based. The distinction has implications for psycholinguistic theories of the way that words are processed by the mind/brain and the way that language processing develops in children or is disturbed by brain damage (see sections 15 and 26).

The properties of Latin that we have briefly sampled here are what lead to its exemplifying the class of inflectional languages, and a fourth class of language, often added to the traditional typology, is the class of polysynthetic languages.

This class is illustrated by Chukchee (also spelled Chukchi), a language spoken in NE Siberia. In (125b), we see a word which corresponds to the phrase in (125a): (125) a.

nəteŋqin

ŋelgən

good

hide

‘a good skin, hide’

b.

teŋŋelgən

‘a good skin, hide’

In (125a) teŋ is the adjective root and nə-…-qin combines with this to form an adjective nəteŋqin ‘good’. In (125b), the adjective root has formed a compound with the noun ŋelgən ‘hide’ to make a single word. There are various ways in which we can show that this is a single word and not just a closely knit phrase, one of which is the fact that adjective roots like teŋ never appear without their prefix nə- and suffix -qin except in compounds.

In (126), we see a similar phenomenon:

(126) a.

tə-lʔu-gʔen

ŋelgən

I-saw-it

hide

‘I saw a/the hide’

b.

tə-ŋelgə-lʔu-k

I-hide-saw-I

‘I saw a/the hide’

Morphology across languages

161

In (126a), the verb form təlʔugʔen has a prefix tə- marking a first person singular subject (‘I’) and a suffix -gʔen marking a third person singular object, agreeing with the direct object ŋelgən ‘hide’. In (126b), three things have happened. Firstly, the object has now joined the verb and formed a compound verb stem ŋelgə-lʔu

‘hide-saw’. Secondly, in so doing it has lost the -n, which in fact is a case suffix.

Thirdly, the verb now ends in a suffix referring again to the first person singular subject. This suffix occurs with intransitive verbs in Chukchee, but this is explicable as the verb in (126b) is intransitive. This is because its original object has actually formed a compound with it (to have this compound appear with an object would produce a structure equivalent to the English *I saw the hide the tent with too many complements).

Compounding of this kind, functioning as an alternative to a syntactically

formed phrase, is known as incorporation. The noun incorporates its adjective in (125b) and the verb incorporates its object in (126b). Adjective incorporation is not very widespread (though in Chukchee itself it is extremely common), but

object incorporation or noun incorporation is very frequently found in the world’s languages. In fact, in Chukchee, object incorporation can apply to the result of adjective incorporation:

(127) a.

tə-lʔu-gʔen

nəteŋqin

ŋelgən

I-saw-it

good

hide

‘I saw a/the good hide’

b.

tə-lʔu-gʔen

teŋ-ŋelgən

I-saw-it

good-hide

‘I saw a/the good hide’

c.

tə-teŋ-ŋelgə-lʔu-k

I-good-hide-saw-I

‘I saw a/the good hide’

Here, we first incorporate the adjective into the noun in (127b). Then, this compound noun, which functions as an object in (127b), is incorporated into the verb in (127c). Words like təteŋŋelgəlʔuk are not especially uncommon or exotic in Chukchee.