Epilogue
At the beginning of this volume we identified as the most important, central objective of this book the search for cross-linguistic associations between word-formation processes and/or the individual parameters of word-formation processes. A number of associations were identified, but the picture obtained is not very optimistic for those who may have expected a solid system of regular distribution of word-formation processes depending on genetic and/or morphological characteristics of languages.
Word-formation is an inherent feature of every language. This is both trivial and crucial. A language can only exist if it is able to give names to new objects invented, discovered, designed, encountered, obtained, produced, etc. by the members of a particular speech community. Exclusive use of borrowings and/or descriptive/analytic devices may be viewed as a symptom of a serious disease in a language, as has been confirmed for Bardi where, according to Bowern (pers. comm.), there is not much in the way of productive word-formation. This might be partly because of language death (there are only twenty-five speakers), but a lack of productive derivational morphology has also been noted for some other Australian Aboriginal languages.
There is no language without concatenating word-formation. This has confirmed one of the fundamental postulates of Natural Morphology concerning constructional iconicity as a most natural way of forming new words. This is also in accordance with the fundamental principles of Marchandean and Dokulilean approaches to word-formation, according to which an object is first conceptually processed as a member of a larger class of similar objects and, subsequently, is distinguished from them by highlighting one of its most characteristic features.
Suffixation and compounding, in this order, prevail, and reduplication and prefixation are frequent in the languages studied. Prefixal-infixal derivation, stem consonant alternation and stress shift are rare. Agent formation prevails over other semantic categories, e.g. patient. Diminutives prevail over augmentatives and feminines over masculines. Certain associations can be found between the above and the independent variables like language family, morphological type or word order, but individual word-formation processes and their features depend more on their genetic rather than morphological characteristics.
Languages may be divided into derivationally rich and derivationally poor. Derivationally rich languages make productive use not only of the major but also of the minor word-formation processes. Typical derivationally rich languages are, e.g. Clallam, English, Hebrew, Ilocano, Indonesian, Karao, Konni, Nelemwa, Marathi, Slovak and Totonac. Derivationally poor languages are those whose word-formation capacity is restricted to a minimum (Bardi, Cirecire) and those which make use of a limited number of word-formation processes, like Kalkatungu, Kwakw’ala, Lakhota, Tatar, Tzotzil and West Greenlandic. From this list it follows that at the poor end of the scale there are mostly agglutinative and polysynthetic languages, the latter exclusively from North America. The rich pole of the scale is fairly heterogeneous. These two groups of languages represent opposite ends on the word-formation richness scale. The other languages can be classified between them in terms of the number of processes used and of their productivity. From the point of view of various morphological classifications of languages, it is understandable, as there are no pure types and, thus, one cannot expect that word-formation features of these languages will be homogeneous.
A sample of fifty-five languages lends itself well to the identification of cross-linguistic default values, but it is not possible to draw any generalizations for families represented by a minimum number of languages. Why certain processes or categories prevail, as has been presented here, is hard to interpret. It may well be that there is not just one explanation, and an interplay of factors gives rise to the use of the same resource in a number of respects. From the perspective of this book, it is our belief that future effort should be guided in three major directions:
(a) an increase in the total number of languages sampled in order to confirm or improve the precision of the findings of this book,
(b) a search for associations between word-formation processes inside selected language families and/or genera represented by a sufficient number of languages, and
(c) as the questionnaire used covered only selected areas of word-formation, an extended scope of the word-formation characteristics covered.
A fascinating feature of any cross-linguistic research is that it makes it possible to reveal the wonderful diversity of individual phenomena. This has been shown, as we believe, in the examples cited in the volume. This diversity ranges over individual languages, families and morphological types. This diversity is a source/reason?/explanation? of the existence of fuzzy boundaries between the individual word-formation processes and also between inflectional morphology and syntax and derivation. In spite of this, the diversity constitutes the unity of the individual phenomena and processes through their most frequent cases that tower over all the less frequent manifestations.
This book draws generalizations based on empirical data, points out associations in word-formation processes cross-linguistically, demonstrates the multifarious manifestations of broadly defined word-formation categories in the languages of the world and illustrates the observations with examples. The aim was to instigate even more extensive and more comprehensive research in this intriguing area of language universals and typology, a task which is insurmountable for one or two linguists. By implication, since it was just the first, tentative probe, many more interesting findings remain to be discovered.
