Inflectional and/versus derivational morphology: clear-cut types or continua?

ABSTRACT

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 of 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 Russell (1998: 128-30), 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 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 3rd 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 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 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 of 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 Matthews' (1983) example, 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 'My (boy) friends have arrived': Temos ESTUDADA a licao '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., Tem ACEITADO todas as nossas propostas 'They have accepted all our proposals' versus As nossas propostas nao foram ACEITAS 'Our proposals were not accepted'. Modern Castilian seems to be explicit enough in pointing to the participial or the adjectival status of the predicate: preparado is participle of PREPARAR in Todas las estudiantes han PREPARADO sus ejercicios '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 estan PREPARADAS '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 [13:nt/13:nd] vs. my learned colleagues [13:nId]: 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--freeom, 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, and similar, under one entry, while divert--diversion, happy--happily--unhappy, etc., are listed separately. There are also scholars who even think that 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. (6) 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: (7) 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, Matthews (1983: 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 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 Hill (1958: 168-70) 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' ..." (Matthews 1983: 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} (exemplified by hotter) and {er/or} (as in teacher, sailor). Specific detailed studies, for example Stekauer'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 sequence/ian/, namely, to try to answer the question whether this sequence was derivational or inflectional, and why the /i/ was finally lost in Middle English. (8) Also such cases as elocution must be paid attention to: namely, while linguists would take it, probably unanimously, as a derived complex form, alongside with a great number of others of the same type (e.g., generation, impression, diversion), the fact is that Modern English has not the expected base *elocute in stock. And, evidently, the form elocution cannot be based on the listed entries, such as elocutionary, elocutionalist: these themselves must be derivatives of a "shorter" form, which, as we can witness, does not exist any longer. Of course, in other cases it does exist, as in COMPUTE [right arrow] computer, computerize, computation. Moreover, semantics of members of the pairs discussed need not be identical: suffices to consider the meanings of, for example, selection and election. If I may use Matthews' arguments (1983: 52-54), "selection" presupposes, and in fact is, a collection of objects which will be or have been selected, but "election" is not equally a collection of people who have been elected! And indeed, the senses of verbs and their derived nouns do not always correspond precisely: while the meaning of direct is clearly comprehensible from He was going in the opposite direction, the phrase the salutation of officers is not at all parallel to something like You must salute officers. Nevertheless, do we understand it in such a way that the meanings of (derived) nouns and their (base) verbs are semantically unrelated? Of course not. But it is also true that in a certain period of the language evolution the relationship between the noun and its verb-base is predictable although in future that of the noun may branch out independently. Discussing this issue we do not necessarily need to probe into such delicate cases as the status of {-sc-}, one that is frequently used to illustrate linguistic change: students in Romance languages will understand that the consonant group {-sc-} must have been inflectional first (IE.*-e-sk-), then it became the signal of inchoativeness (Gr/L yiyv[omega][sigma][kappa][omega]/co-gno-sc-o, meaning 'I came to know'), and in the end it is inflectional again (as in Sp. conozco). All advanced learners of English will understand the meanings of, for example, intrepid or trepidation although there is no *trepid now as their base form. They will also know what butcher is, even though the form butch, formally parallel to teach--teacher, do--doer, and the like, has nothing to do with the common meaning. And, last but not least, a scholar can meditate over the meanings, or over differences of meanings in such pairs or even such concatenations of expressions as warm--warmth, idiom - idiomatic--idiomatically: stupid--stupidity--stupidness (?): profound--profundity--profoundness, and the like.

However brief my arguments may have been, I hope to have shown that the issues concerning what is generally referred to as derivation, by itself as well as in relation to inflection, are not easy to tackle and brought to a definite end, unless we accept the idea of "spectrum" rather than clear-cut, box-like categories.

Let us return to the problem again now, assuming that the derivational morphology does exist, although its boundaries may be shady, showing certain overlapping characteristics. With reference to what I have been trying to elucidate, searching for relevant criteria, I believe that there must be two types of derivational morphemes, namely, (1) those that change the syntactic category of the base morpheme to which they apply (e.g., read--readable), and (2) those that do not (e.g., do--undo). Yet it always holds that derivational processes create new naming units, or lexical items. The criterion to supplement that of the syntactic component involved is a semantic one: the meaning of the affix is believed to alter or to affect the meaning of the base morpheme to a certain extent, which varies from case to case. As for example Bybee (1985: 82) maintains, "in case of verbs, it means the extent of which the meaning of the morpheme affects the description of the situation, and for nouns, the extent to which the meaning of the morpheme changes the referent of the noun". And indeed, if we ask the question whether the following derived forms differ in meaning from the base ones, we have to give a positive answer: untie is just the opposite, distinct event as compared to tie; unhappy is not altogether the same as sad. (9) baker refers to a person, linguistically an agent of the activity described by the verb BAKE; professorship denotes a status whereas professor is the bearer of the status. On the other hand, the same type of reflections on the suffixes {ly} or {ing} would very probably make us admit that the expected change in meaning is almost none. Namely, the quality expressed by the adjective such as diligent will not change after -ly has been attached to produce diligently; gerundial nominalisations in -ing, e.g., reading, can appear in syntactic positions from which finite verbs are excluded, but they do not change the situations described by the underlying verbs. Thus the interpretation of read in bed and reading in bed will not differ much in, e.g., John reads in bed and Reading in bed is fun (Bybee's examples, 1985: 83-85). By projecting the findings on to the issue of category changes, I arrive at the following conclusion: the derivatives which bring about a considerably different meaning will not change their syntactic categories as compared to those of their bases (type: untie, unhappy, professorship); and vice versa, the derivatives which add just a little or next to nothing to the meaning of their bases will tend to change their categories (type: diligently, reading). From the other way round, what plays a very important role in these considerations is the amount of semantic change as made by the application of the affix to the base. We can observe that the greater is the difference between the meaning of the derived form and the meaning of the base, the greater is the likelihood that the affix is truly derivational. One would rather think not in terms of two (different) types of derivational morphemes but in terms of a continuum, which will, easily and satisfactorily enough, embrace all the complex forms discussed here above.

It would be a plain sailing from now on if we could pronounce that the desired continuum were extended in both directions, namely, to primary lexical units as well as to inflected forms. And indeed, this seems possible, provided, of course, that we respect the functioning of language in its miraculous complexity. The distinction between a primary word and a derivative is easy to follow since either is represented in the lexicon and the meaningful difference between them is a matter of synonymy, with all consequences involved. Thus the derivative largeness seems to be synonymous to size, and so also unhappy and sad. The difference, and the choice, between each two of them lies in the fact that the content of largeness does not cover the whole content of size, and that unhappy is not altogether identical with sad. And yet, when construing a sentence the speaker has to consider the balance between the semantic content and the grammatical form. Namely, when the speaker's intention is not to deliberately point out the difference in meaning between the base and the derivative, in other words, the difference which is conveyed by the affix, then the primary, monomorphemic expression will be preferred. Therefore, e.g., size or sad, respectively, will be the first at hand, while largeness or unhappy will be used only intentionally. We can only marvel at the fact that syntactic categories of the derivatives will match those of the potential monomorphemic primary lexical equivalents: both size and largeness perform the function of nouns, and both sad and unhappy behave as adjectives. Viewed from these broader aspects, it does not seem to be as much important to deliberately look for parallels which exist between the amount of meaning conveyed by the affix as compared to the meaning of the sole base, on the one hand, and the category change, on the other hand. This is not even possible, also because we are unable to define a unit in order to measure the amounts of the meaning added. Nevertheless, the idea of distinctions between respective derivational morphemes being viewed as gradual and representing a continuum is welcomed.

What remains to discuss is the sphere of transition between derivational morphemes and inflectional ones. Like some of the derivational morphemes, those of the type {ly} and {ing}, inflectional morphemes add no lexical meaning to the base. In other words, while certain morphemes add some (at least little) amount of lexical meaning and are for that reason considered derivational, inflectional morphemes are those which leave the lexical meaning of the base untouched. What makes it possible for them to be referred to as genuine morphemes is the grammatical meaning they convey. Inflection, that is to say, is required by the syntax and it is derived by rule, not in the lexicon. As Bloomfield (1933: 223) has it, thinking of highly structured sets of expressions of regular patterns, inflection is characteristic of a "rigid parallelism of underlying and resultant forms". Thus, for example, nearly all nouns in English underlie a derived plural, all verbs underlie a derived past tense, and the like. The choice of inflectional morphemes is therefore obligatory for the speaker, which is not, in this sense, the case of derivational morphemes. This obligatoriness is likely to be the only relevant criterion for the possible distinguishing between inflectional and derivational morphemes.

It follows from what has been claimed that a speaker makes, first of all, an arbitrary decision on lexical meanings. And this means that derivational morphemes, if such a choice of alternative complex forms is the speaker's intention, are attached closer to the root. Only after this decision has been made, requirements on syntax, namely the choice of appropriate inflectional morphemes, will be satisfied, and that obligatorily. And here lie crucial points of my way of viewing the issue, which I am now going to present as a conclusion.

I do not see any direct relationship between syntactic categories and the inflectional and derivational categories. Changes in syntactic categories are only made within the sphere of derivation! Or, more precisely, changes in syntactic categories are frequently made within the sphere of derivation if the semantic difference between the given members of a derivational paradigm is only a slight one. In other cases, namely, if the complex member of the opposition has to be chosen, since different semantics is needed, no change of syntactic category takes place. At least one example to either "model" situation will illustrate, to supplement those I discussed above.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

As I have also pointed out, we can hardly measure the amount of semantic change. There is no unit available to tell how much friend--friendship differ as compared with friend--friendly or friendly--unfriendly, or perhaps hot--hotter. But we somehow sense the extreme values, allowing for a space in between for a cline-like continuum. With the maximum semantic change syntactic categories have been shown as untouched: yet these may remain the same even if the semantic difference draws to zero. If this is the case, inflectional morphemes/morphology/categories, and the like, are referred to. Put the other way around, certain morphemes are attached to the base not because they are required by meaning but rather by needs of grammatical concord and government. Were it not for inflectional languages, or should we apply a synchronic analysis to English only, the distinction between "inflectional" and "derivational" would be of but a little use. Strictly speaking, there is hardly any other morpheme in Modern English but {s} attached to verbs and required merely by grammar. With others, including {s} in noun plurals, the choice may be explained as arbitrary. Indeed, I call and I called are two different "things", as well as boy and boys are two different "things", and so also the above discussed hot and hotter. Objections can be raised easily, of course, but acceptability of this view can only hardly be denied absolutely. What I claim, quite overtly, is the fact of a sort of "fluidity" between the traditionally acknowledged inflection and derivation. I hope to have depicted the sphere of derivational morphology as a continuum: now, in order to allow for certain universality, I do not turn down the existence of inflection, but, at the same time, I propose projection of the continuum on to the inflectional morphology.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

This approach will lift from us the burden of a precise defining of respective morphemes and makes it easier to treat such (specific, debatable) cases as I exemplified, be it {ing}, {ed}, {er/or}, {ly}, or {dom} complex words.

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

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

(3) Mind the syntactic orderings and the different stress layouts within the NP to bring about different meanings: a [sup.1]young dancing [sup.1]master vs. a [sup.1]dancing [sup.1]young [sup.1]master.

(4) I am grateful to P. Stekauer 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.

(5) Here thought is a form coming into existence through a derivational process, unlike the homonymous past-tense form.

(6) In the Longman Dictionary of Contemporary English. Longman Group, Ltd., (1978).

(7) For instance, Webster's New World Dictionary (1989).

(8) Arguments are found in F. Colman's (1996) study. See References here after.

(9) Here I do not fully agree with Bybee (1985) in that the result of a derivational process has a lexical monomorphemic item as its counterpart, at least not from the meaningful point of view.

REFERENCES

Archangeli, Diana--D. Terence Langendoen (eds.) 1998 Optimality Theory: An overview. Oxford: Blackwell Publishers, Ltd.

Bloomfield, Leonard 1933 Language. New York: Holt, Rinehart and Winston.

Bybee, Joan L. 1985 Morphology: A study of the relation between meaning and form. Amsterdam--Philadelphia:

John Benjamins Publishing Company.

Colman, Fran 1996 "Morphology: Old and Middle English--derivational and inflectional", in: Jacek Fisiak (ed.), 3-28.

Fisiak, Jacek (ed.) 1996 Middle English miscellany. Poznan: Motivex.

Hill, Archibald A. 1958 Introduction to linguistic structures. New York: Harcourt, Brace & World.

Kavka, Stanislav 2000 "Some hints on the importance of teaching idiomaticity", Studia Anglica Posnaniensia 35: 205-216.

Matthews, Peter H. 1993 Morphology. Cambridge: Cambridge University Press.

Russell, Kevin 1998 "Optimality Theory and morphology", in: D. Arehangeli--D. T. Langendoen (eds.), 102-133.

Stekauer, Pavol 1998 An onomasiological theory of English word-formation. Amsterdam--Philadelphia: John Benjamins.

STANISLAV J. KAVKA

University of Ostrava