Page 298 - DENG504_LINGUISTICS
P. 298

Linguistics



                  Notes          Neo-classical roots such as scope and graph can also be used nowadays as words, but in that case they
                                 have a more specific meaning than in these compounds. Such non-lexical roots are called combining
                                 forms since they only occur in combination with other morphemes. These bound roots cannot be
                                 considered affixes since that would imply that words such as necrology would consist of affixes only.
                                 This goes against the idea that each word has at least one stem. Thus, we might adapt our definition
                                 of what compounds are, and define them as combinations of lexemes and/or non-affixal roots.
                                 The bound morphemes in neo-classical compounds have an identifiable meaning, but there are also
                                 morphemes that have no clear meaning. In the word cranberry the part berry is identifiable, and this
                                 makes us interpret the word cranberry as denoting a particular kind of berry. Yet, cran- has no particular
                                 meaning. Similarly, the Dutch compound stiefvader “stepfather” denotes a particular kind of father,
                                 and hence can be parsed into stief and vader. However, the morpheme stief does not occur as a word.
                                 This phenomenon of cranberry morphemes is widespread, and is to be expected since complex words
                                 can lexicalize and thus survive, even though one of their constituent morphemes has disappeared
                                 from the lexicon. The following examples from Dutch illustrate the same phenomenon for derived
                                 words with suffixes that are still used for coining new words (the constituent before the suffix does
                                 not occur as a lexeme):

                                 (7)  arge-loos “naive”, beslommer-ing “chore”, dier-baar “dear, precious”, le-lijk “ugly”, moei-zaam
                                      “difficult”, sprook-je “fairy tale”, veil-ig “safe”
                                 These recognizable suffixes determine the syntactic category of the word of which they form a
                                 constituent. For example, -baar is a suffix that creates adjectives, and hence dierbaar is predictably an
                                 adjective. This implies that when we have to decompose words into morphemes, not all morphemes
                                 have an identifiable lexical or grammatical meaning. Cranberry morphemes like English cran- and
                                 Dutch dier- thus form a problem for an exclusively meaning-based definition of the notion morpheme.
                                 This also applies to another kind of non-affixal bound root, the recurrent constituents of words
                                 borrowed from Latin such as the following English verbs:

                                 (8)  conceive, deceive, perceive, receive
                                      adduce, deduce, induce, produce, reduce
                                      admit, permit, remit, transmit

                                 It makes sense to consider these words complex, because of recurrent elements such as ad-, con-, de-,
                                 in-, per-, pro-, re-, and trans- which are prefixes, and bound roots like -ceive, -duce, and -mit. Although
                                 these bound roots have no identifiable meaning, they should be recognized as mor-phemes since
                                 they determine the form of corresponding noun: all verbs in -ceive have a corresponding noun in -
                                 ception, those ending in -duce one in -duction, and verbs in -mit one in -mission. There is a wealth of
                                 such bound morphemes in the non-native part of the English lexicon, as the following examples
                                 illustrate:

                                 (9)  arct-ic, cred-ible, in-del-ible, gradu-al, mor-al, mus-ic, negoti-ate, per-for-ate, per-nic-ious
                                 In lexeme-based morphology these bound roots do not have a lexical entry of their own, they only
                                 occur as part of established (listed) complex lexemes. In morpheme-based morphology, on the other
                                 hand, they will have to be represented as bound lexical morphemes with their own lexical entry. The
                                 advantage of the lexeme-based approach is that it correctly predicts that new combinations of a
                                 prefix and a bound root such as demit or perduce are not to be expected, because we cannot assign a
                                 meaning to such new combinations.
                                 Boundness of morphemes is also created through allomorphy. Allomorphy is the phenomenon that
                                 a morpheme may have more than one shape, corresponds with more than one morph. A morph is a
                                 particular phonological form of a morpheme. Allomorphy is found in both affixes and root morphemes.
                                 In the Italian examples in (4) we saw the prefixes al- and ap-. In fact, these are two allomorphs of the
                                 prefix ad-, in which the final consonant /d/ has assimilated to the first consonant of the root morpheme.
                                 This kind of allomorphy can be accounted for by assuming one common underlying form /ad/ for


        292                              LOVELY PROFESSIONAL UNIVERSITY
   293   294   295   296   297   298   299   300   301   302   303