Inflectional vs Derivational Morphology.pdf

INFLECTIONAL AND / VERSUS DERIVATIONAL MORPHOLOGY:

CLEAR-CUT TYPES OR CONTINUA?

Some colleagues of younger generation may accuse me of being too faithful to traditionalism in

observing linguistic phenomena. May I argue, frankly indeed, I do appreciate the contemporary modern

computerising trends in collecting and analysing language data; yet I still believe in a peaceful meditation as a

counter-balance to technical-like approaches, finding the results fairly efficient and, hopefully, significant

enough for both theoretical and applied linguistics. One of my little ideas is to show that the traditionally as well

as newly established categories, however much useful notions these may be, are not clear-cut ones but rather

represent “cardinal”, i.e. supporting, even extreme points in a spectrum 1 . All kinds of transitions are observed,

the properties of respective categories oscillating around the hopefully well-defined points and merging into one

another. Examples can be found in all spheres of language analysis, from phonetics to semantics, and so also in

morphology.

Any reader linguistically oriented will have noticed that there have been scholars who

zeroed in on morphology as well as those who doubted, as it were, the place of this discipline

in the overall organisation of grammar. The question reads whether there is any need for a

truly morphological component in language at all. Does anything like a morphological

component exist, in the sense of a parallel to what everybody generally acknowledges as a

phonological and a syntactic component – this is the question that some linguists tackle. TG

grammar proponents, for example, are those who dispute the existence of morphology,

claiming that all the work of assembling words can be performed by principles of semantics,

syntax and phonology. Representatives of so-called ‘lexical hypothesis’, or, on the other hand,

those who believe in omnipotence of syntax in word-formative issues, they all seem to be

convinced that they got rid of the problems involved, the problems which exist objectively all

the same, irrespective the satisfaction-giving oversight. I have been looking for a resolution of

some of the many remaining issues, believing, of course, in the existence of morphological

component in grammar description. To my knowledge, the discussions are still far from being

finished; even the recent Optimality Theory seems to fail in bringing a satisfactory agreement

on how the morphological component interacts with the other components. Following K.

Russell (1998: 128 ff.), for example, one of the unresolved questions, which depend largely

on the overall approach to grammar and to morphology in particular, is the difference between

inflection and derivation (or, word-formation). In other words, the issue reads whether

inflectional morphology and derivational morphology operate on different principles, namely,

whether the two types of morphology, as distinguished traditionally, are fundamentally

different. Or do we only look for problems where there are none?

In the following I am going to opt in favour of keeping inflectional and derivational

morphology apart, although, as I will try to argue, these cannot well be viewed as clear-cut

categories, or types. The reader will, hopefully, understand that the feasibility of one type

being “drifted over” to another is not the same as considering the whole morphology to be

either fully inflectional or fully derivational, i.e. lexical. As a matter of fact, what I take into

account is the functional characteristic of formatives, or morphemes, these being referred to as

inflectional or derivational, respectively. Yet it would be wrong, as I argue, to assume that

there are languages (English included) the morphemes of which are exclusively inflectional or

exclusively derivational; there are, no doubt, morphemes that are endowed with either

function, in dependence on the type of operations they enter. Here the interaction of the

morphological component with the components of syntax and even semantics must be taken

into account and explained in a satisfactory way. Complex evaluation of individual, yet

I have been trying to do so in courses of lectures, recently also on issues of idiomaticity. See Kavka, S. (2000).

sometimes prototypical instances is always expected; as elsewhere, it holds that no single

criterion should be applied in absolute terms.

Let us begin with general cases of interaction, first of all with the syntactic

component . Here, as is believed, inflectional morphemes are those that are required

obligatorily by the sentence syntax; for good measure, they enter operations which leave

syntactic categories of base morphemes untouched. For example, Latin mons/montem

[‘mountain’] is a noun, bonus/bonum/bona [‘good’] is an adjective, amo/amare/amavit [‘to

love’] is a verb, etc. Yet it is true that these convey some more “elements”, namely

grammatical meanings, such as gender, number, tense, etc.; these are labelled traditionally as

‘grammatical categories’, i.e. properties which may be syntactically relevant in fitting the

word in a sentence. So also in Modern English: in cloud/clouds , for example, the morpheme

{s} of the latter member conveys the grammatical meaning of number. Hence, it is expected,

the two members of the inflectional category always belong to the same syntactic category;

namely, both act as a syntactic subject or an object and they are traditionally referred to as

nouns. But is this {s} in the English clouds selected really on the same grounds as the Latin

{em} or {a} or {v-it} in the aforesaid examples? In other words, can the English {s} in

(syntactic) nouns be defined unanimously as an inflectional morpheme?

There are also morphemes of different syntactic functions, though. As a matter of fact,

the choice of these is not determined by syntax but, on the contrary, once they are chosen, the

new, complex term requires being treated syntactically in a different way. Thus the Latin

erogo [‘I pay’] is a verb, whereas erogatio [‘payment, distribution’] is a noun, and the English

cloudy is classified as an adjective, in opposition to cloud as a noun; likewise, enrich is a verb

while the base rich is an adjective. Apparently enough, the members in opposition do not

belong to the same grammatical category. In other words, the morphemes {atio]; {y}, {en} in

the present examples enter so-called derivational operations, through which new naming units

arise, belonging to syntactic classes different from the base. As is commonly known,

derivational morphemes have been referred to as affixes, and more specifically as suffixes,

prefixes and infixes, in dependence on their position towards the base.

If explained in a simple way as above, the distinction between inflectional and

derivational morphemes could be understood with only a modicum of difficulty. Yet why do

we find among linguists different approaches to the distinction and why none of them seems

to have ever provided a unanimous discrete division between the two categories of

morphemes, and consequently, between inflectional and derivational morphology? I myself

am doubtful of the simplicity: the matters seem to be so neatly organised that they cannot be

true. A few examples will illustrate some problems involved.

One and the same morpheme can be regarded as inflectional in one case and as

derivational in another case. Thus the English {er} attached to the base big leaves the

resultant complex form syntactically untouched (i.e. bigger as well as big are both adjectives

of the same syntactic behaviour), whereas having been attached to teach , resulting in teacher,

it entails a change from a verb to a noun, which brings about syntactic consequences (i.e.

teach can only be a predicate, teacher performs the functions of a subject or an object).

Similarly, the morphemes {ing} and {ed} can also fall in either category: since they are

attached mostly 2 to verbal stems, one is tempted to take them as inflectional, likewise {s} and

{ed} marking the 3 rd person singular present and the past tense, respectively. (Here also forms

like wrote, thought, written , etc. belong, treated by some linguists as coming into existence

Not always are the bases primary verbs, e.g. skilled , or naked , viewed synchronically.

through a ‘process’ change.) And indeed, dances/danced/dancing in She dances in a ballet

group.; She danced with John only.; She’s been dancing all night. are all verbs acting as

syntactic predicates. In She fell in love with a young dancing master. , however, the form

dancing will hardly be described as Present Participle of DANCE but rather as (homonymous)

Adjective DANCING . The change of its syntactic category in this instance speaks in favour of

treating {ing} as a derivational morpheme. 3 Past participle forms such as in heated, crowded,

written, bent , and the like, are also worth a closer inspect. Following the arguments by P.H.

Matthews (1993: 55-56), one can only with difficulty decide on one or the other category of

these. If we can rely on the casual test of ‘very’-intensifier modification, then the form heated

remains the Past Participle HEAT (since not * a very heated room but rather a well heated room

is acceptable), and crowded should be a derived Participial Adjective CROWDED (thus

allowing for a very crowded room ). 4 With the form written the situation is yet more blurred:

The contract was written points to the verbal status of written << WRITE , this being a passive

construction rather than the equivalent of ‘ The contract was a written one .’ (i.e. not made, for

instance, by word of mouth). On the other hand, its occurrence in the Attributive position,

namely as written << WRITTEN , is rather restricted: one would expect to say a hand-

written/type-written contract , not *a written contract , even though there are no verbs *HAND -

WRITE, *TYPE-WRITE . It seems useful here to apply one more factor, namely, to consider the

linguistic change: then ‘written’ is diachronically based on its verbal stem, together with

others that are called Past Participles, but viewed synchronically its status can in certain cases

be arbitrarily adjectival. In other cases, like bent , it is not always certain which category the

respective forms belong to. A bent pin (to use P.H. Matthews’ example; op.cit.) may be

understood either (1) as a pin that has been bent (maybe on purpose, maybe incidentally, and

thus damaged and of no use) or (2) as a specific kind of pin that has a certain shape (like we

have other types of instruments, e.g. safety-pin, drawing-pin, or clasp-knife). Most of us will

probably agree upon bent in a bent pin (1) as being Participle of BEND and thus belonging to

the inflectional category. With bent in a bent pin (2) the decision is not so unambiguous,

although we would not be strictly against its Adjectival status. Moreover, there is a genuine

Adjective BENT , as in a bent copper (i.e. ’dishonest’) or in Jim seems to be bent on becoming

a musician (i.e. ‘determined to become’). It is because here bent enters certain types of

idiomatic expressions: a bent pin (2) is felt as a more cohesive expression than a bent pin (1),

the latter being a free combination, as also are a bent nail, a broken leg , and similar.

The examples above must have shown that the decision on the inflectional or the

derivational status is not a matter of clear-cut, precisely defined boundaries. There are cases

where the shift from Participle to Adjective takes place. Minimum of morphological signals in

Modern English to show grammatical concord makes the situation less apparent than it was in

Latin and than it is in its descendant idioms. A student in Romance languages will have

noticed that in Italian and, in certain cases, in Portuguese, too, participial forms in predicate

agree with their subjects or objects, e.g. I miei amici sono ARRIVATI . [i.e. ‘My (boy) friends

have arrived.’]; Temos ESTUDADA a lição. [i.e. ‘We have finished studying the lesson.’].

Moreover, in Portuguese participial doublets of certain verbs appear in one or the other form,

the selection being due to the grammatical voice used in respective cases, e.g. Têm ACEITADO

tôdas as nossas propostas [i.e. ‘They have accepted all our proposals.’] versus As nossas

propostas não foram ACEITAS . [i.e. ‘Our proposals were not accepted.’]. Modern Castilian

Mind the syntactic orderings and the different stress layouts within the NP to bring about different meanings:

a Ç young dancing Ç master vs. a Ç dancing Ç young Ç master .

I am grateful to P. Štekauer who, having read the first draft of this article, reminded me, besides others, of the

fact that the “very-test” would not work with many genuine adjectives (e.g. round, weekly, blind, dead ) since not

all adjectives (and likewise nouns, verbs, etc.) are always carriers of characteristics of their respective class.

seems to be explicit enough in pointing to the participial or the adjectival status of the

predicate: preparado is Participle PREPARAR in Todas las estudiantes han PREPARADO sus

ejercicios. [i.e. ‘All the (female) students have prepared their tasks.’], whereas the same form

is Adjective PREPARADO , which takes on respective inflectional morphemes to meet the

requirements of concord: for instance, Todas las estudiantes están PREPARADAS . [i.e. ‘All the

(female) students are prepared/ready.’].

It is worth of note that something similar, though only partially and without a hundred

per cent regularity, takes place in English, too. What I have in mind is some irregular verbs,

e.g. He has learnt / learned [lз:nt / lз:nd] vs. my learned colleagues [lз:n ö d] ; He has drunk a

few too many. vs. a drunken party; He has lit a match. | The match is lit. vs. a lighted match;

He was bereaved of all hope. | He is bereft of his senses. vs. Poor bereaved mother. Out of

others, at least the following may be taken into account, the latter in each pair being defined

as Adjective in English dictionaries: born / borne, hung / hanged, melted / molten, proved /

proven, shone / shined, shaved / shaven, shrunk / shrunken, sunk / sunken, struck / stricken,

sunburned / sunburnt.

On the other hand, the examples such as generate-generation, decide-decision, think-

thought 5 , respectively, will be looked upon as not only different forms but also different

naming units. This view is based on the fact that on the syntactic level the resultant complex

forms as nouns behave differently from the base morphemes as verbs. The list can be

extended by (if not all, then at least some illustrative) examples in which the affixes are

believed to be productive: cloud-cloudy, rich-enrich, warm-warmth, large-largeness,

productive-productivity, diligent-diligently, read-readable, etc. Here, too, members of the

respective pairs do not belong to the same syntactic categories, which makes us speak in

terms of derivation. This is not always the case, though: both neighbour and neighbourhood

are nouns, and so also friend and friendship, king and kingdom ; and both will perform

identical syntactic function. Yet hardly anybody would ever think in terms of inflection! Does

it mean then that all the three morphemes, i.e. {hood}, {ship}, {dom}, are derivative ones?

Such a pair as free-freedom will immediately cast doubts onto our mind. How come, one must

ask, that one and the same morpheme attached to the base brings about a change in the

syntactic characteristics of the complex form in one case (e.g. free-freedom, bore-boredom,

official-officialdom ), and leaves it untouched in another case (e.g. king-kingdom, duke-

dukedom, professor-professorship )? This is something that most (English) linguists take for

granted, referring to “sub-categorisation” of the base through a suffix. Yet again, the idea of

less rigidly defined clear-cut categories comes to mind, as well as the risk of applying one

single criterion in absolute terms. A few more examples will demonstrate that no change in

the syntactic category takes place and yet we are sure of derivation: (to) tie-untie, happy-

unhappy, loyal-disloyal, symmetrical-asymmetrical. Willingly or unwillingly, we have to

admit that the syntactic criterion alone is not sufficient, and that its application as made so far

will not bring about satisfactory results. Therefore semantic characteristics must also be

taken into consideration.

Even a brief, momentary look into dictionaries will make us believe that

lexicographers are sure about the issues that I have been discussing. We find the forms such

as book-books or write-wrote-written 6 , and similar, under one entry, while divert-diversion,

happy-happily-unhappy , etc. are listed separately. There are also scholars who even think that

Phrasal and prepositional verbs are listed separately, though.

Here thought is a form coming into existence through a derivational process, unlike the homonymous past-

tense form.

arguments adduced by lexicographers are something brought to us on tablets of stone; then,

no wonder, dictionaries must give unambiguous data. However, lexicographers ARE aware of

those many fuzzy points; yet the purpose of a dictionary does not allow for the discussion I

am offering in the present article. An inquisitive reader would certainly ask questions why, for

instance, compress-compressible-compression-compressor are listed as individual, separate

entries, whereas compressibility is marked as ‘-bility’ under the entry com-pres-si-ble. 7 Does

that mean that the ‘bility’ is a different morpheme from, say, ‘or’ ? And, is the ‘bility’ a

genuine morpheme at all? Linguists would rather refer to {ility} as attached to ‘compres-si-

ble’ , or, to be in agreement with the general opinion, they would regard ‘bility’ as

representing two derivational morphemes, i.e. {able} + {ity}. Moreover, native speakers (as

well as advanced learners) would not feel happy of not finding such a derivative (?) as

compressive . Here some dictionaries will offer more satisfactory answers: 8 the bold-typed

entry compressive and the attached compressive - ly will certainly provide the consultant with

more information. Also the entry compressed listed in one but not in the other dictionary is

more illustrative and closer to the real state of matters, in the sense I tried to hint in the

paragraph dealing with so-called participles. The more so, practitioners in English will be

content with finding respective ‘words’ as ones truly existing in Modern English. Technical

ways of listing the respective forms in a dictionary can vary, and as it seems to me, there is no

definite way, in most of the cases, to show whether the given expression belongs to the

inflectional or the derivational category. Therefore I am reluctant to agree with such opinions

as, for example, P.H. Matthews (op.cit.: 45) takes, claiming that “ to say that an opposition is

‘lexical’ means that the difference is stated in the lexicon or dictionary”.

Does it hold (what P.H. Matthews also claims) that an opposition is “…non-lexical if its

terms are in general selected by the grammar” ? In most cases we do find A (simplex) and B

(complex) in an inflectional opposition if the choice between them, in some instances at least,

is determined by a general grammatical rule. Languages that have a (fully) developed

inflectional system can be a good proof of that. Let us return to one of the tricky issues,

though: What about, for example, the forms hot-hotter ? Is hotter a mere inflectional form of

HOT , one that we traditionally call ‘comparative’, or is it a different lexeme HOTTER ? And, as

I have pointed out above, is the morpheme {er} inflectional or derivational? Originally, I

suggested that it might be one of both functions. Yet just in this point linguists’ opinions

differ: hot and hotter are not selected according to a grammatical rule. This is what A. A. Hill

(1958: 168 ff.) says, taking the two forms as two different lexemes, i.e. HOT , and HOTTER .

Most linguists, however, think that this opposition is a matter of inflection because “there is no

indisputably simple form that can be substituted for ‘hotter’…” (P. H. Matthews, op.cit.: 51). Indeed,

we cannot substitute anything grammatically equivalent for hotter ; but we can easily

substitute the complex automation by the simplex beer in, for example, Automation / Beer is

a good thing. But the question remains open all the same: {er} can be shown as the morpheme

performing two functions, or there may be two different homonymous morphemes, namely

{er}¹ (exemplified by hotter ) and {er/or}² (as in teacher, sailor). Specific detailed studies, for

example P. Štekauer’s (1998) onomasiological theory, speak in favour of even more than two

homonymous ‘er’ morphemes!

However, my principal intention is not to discuss individual instances but rather to

grasp general trends and tendencies. And here also other arguments should be elucidated.

Most of them can be understood more easily if viewed from a historical perspective. This

does not mean only to consider, for example, the issue of the Old English phonological

7 In the Longman Dictionary of Contemporary English . Longman Group, Ltd., 1978.

For instance, Webster‘s New World Dictionary . 1989.

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: