Morphophonology

Up to this point we have looked at allophonic variations (contextual variation of the sounds belonging to the same phoneme), which take place within a single morpheme. We should also point out that several of the processes that are shown to be crucial in accounting for the allophonic variations in languages can also be found active across morpheme boundaries. A brief discussion of these is useful, because they are frequent sources of confusion for students. When morphemes are combined to form bimorphemic (with two morphemes) or polymorphemic words, many of the assimilatory phenomena discussed can be present there too. Such things can also manifest themselves when two words are spoken consecutively. What we see in these instances, then, is the contextually determined alternations (different phonetic forms) of a morpheme.

To illustrate what has been said thus far, let us look at the nasal assimilation rule in Italian we cited earlier (in Phonemic Analysis: A Mini-demo). For explicitness, we give the relevant details below.

As we see in the left column, [n] appears before [t, d, s, e] and, in the right column, [ŋ] appears before [k, g]. The sounds [n] and [ŋ] are in complementary distribution, since they share no phonetic environments. The velar allophone, [ŋ], appears only before velar stops, and the alveolar, [n], appears elsewhere. The phonetically motivated (nasal assimilating to the place of articulation of the following segment) contextual variation occurs within a morpheme, and thus qualifies for allophonic variation.

The same/similar phonetically motivated nasal assimilation is also found across morpheme boundaries in several other languages. Observe the following:

All the adjectives listed are preceded by the same negative prefix that manifests itself as either im or in orthographically. As for the phonetic manifestation, we have three forms: [ɪm] for the left column, [ɪn] for the middle column, and [ɪŋ] for the right column. That is, the pronunciation of the negative prefix is different only with respect to the nasal consonant, which is bilabial, [m], before adjectives that start with a bilabial sound (left column), and velar, [ŋ], before ones that start with a velar sound (right column). In other instances, the nasal is alveolar, [n], in the prefix (middle column). This predictable alternation of the nasal is the result of the place of articulation assimilation that is reminiscent of the Italian example discussed above. Although the phonetic motivation (place of articulation assimilation for articulatory ease) of this alternation is the same in these two situations, they are different structurally. While the Italian contextual variation was allophonic in nature (occurring within one morpheme), the case of the English negative prefixes does not deal with allophones of the same phoneme, but rather shows contextually predictable alternations among separate phonemes. That English [m], [n], and [ŋ] are not allophones of the same phoneme but belong to separate phonemes /m/, /n/, and /ŋ/ can clearly be shown by the following triplet: sum [sΛm], sun [sΛn], sung [sΛŋ]. Thus, what is revealed in the case of English negative prefixes is that there is an alternation of different phonemes for the same morpheme (indicators of the same meaning unit). Such cases are traditionally called morphophonemic alternations, and the different phonetic manifestations of the same morpheme (morpheme alternants) are called the allomorphs.

The previous case demonstrated allomorphic alternations across morpheme boundaries in the same word. These cases, however, are not restricted to morpheme boundary situations in one word but can occur across the boundaries of two separate words. For example, if we examine the phonetic manifestations of the morpheme meaning “one” in Spanish, we find the following.

While the morpheme meaning “one” in Spanish is consistently spelt as un, its pronunciation varies among [um], [un], and [uŋ]. Here, again, we have a familiar picture regarding the nasals assimilating to the place of articulation of the following obstruent. This time, however, the allomorphs (phonetically conditioned variants of the same morpheme) of the meaning unit “one” reveal the alternation across two words.

Another possibility is to find some feature-changing assimilatory processes that are restricted only across morpheme boundaries (i.e. acting only as morphophonemic processes), and with no parallels in allophonic processes of monomorphemic words. To illustrate this, examine the following past tense endings in English:

The above examples show that the regular past tense ending in English has three predictable phonetic manifestations. We have [-əd] if the last sound of the verb is an alveolar stop, /t, d/ (schwa insertion before another alveolar stop). If the verb-final sound is not an alveolar stop, however, then the shape of the past tense ending is an alveolar stop, [-t] or [-d], which is determined by the voicing of the verb-final sound. We have the voiceless [t] if the verb f inal sound is voiceless (middle column); however, the form is the voiced [d] if the verb-final sound is voiced (right column). This is a clear case of voicing assimilation. However, this does not mean that this sequencing restriction occurs throughout. While the past tense of the verb ban [bæn] is necessarily [bænd] and cannot be [bænt], this does not mean that we cannot have a final con sonant cluster with different voicing in its members ([-nt] sequence) in English.

Words such as bent, tent, etc. reveal that there is no such restriction within a single morpheme, and that the assimilatory situation is at work only across morpheme boundaries.

We should also mention that while morphophonemic alternations reveal several feature-changing assimilatory (phonetically motivated) processes, they are not limited to those only. Several other processes, such as epenthesis (insertion) and deletion of segments, as well as metathesis (transposition) of segments, can be cited among those.