As has been adverted to in earlier posts, unstressed vowels are often rendered differently in British and American English. The word pentagon in British English is thus heard with two reduced vowels, namely schwas, in its unstressed syllables––[pɛ́ntəgən]—whereas in American only the medial vowel is reduced, hence [pɛ́ntəgòn]. Conversely, British speakers regularly keep the vowel of the constituent {-land} unreduced in pronouncing the American state Maryland, which no American would do. The same distribution applies to most other words containing this constituent, e. g., inland, although free variation is possible in some, like mainland.

Vowel reduction in unstressed syllables is one species of phonological alternation that goes by the name of SANDHI, a Sanskrit term that has been used in linguistic analysis for more than a century. (It is usually pronounced [sǽndi], like the first name “Sandy,” or [sɑ́ːndi], identical with “Sunday” for some British English speakers.) Variations and alternations at the boundaries of constituents is one of the core instantiations of sandhi. Phenomena associated with phonological sandhi rules of this sort have two basic functions in language, one systemic, the other textual. Their systemic function is ICONIC in that they produce distributions of phonetic values in utterances which reflect the distinctive or allophonic value of the features in question and the markedness relations that hold between values of the same feature opposition. Their textual function, on the other hand, works to signal cohesion between elements and is thereby integrative, which is to say that this function is INDEXICAL: it signals, for instance, that constituents of the word in question are connected to each other in patterns of internal cohesion.

The upshot of this semiotic account is to enable the understanding of variation in language as something coherent, not arbitrary. When the constituent {-land} in mainland is treated  (= “understood”) as a bound element subordinate to the meaning of the whole, its vowel loses its stress in the compound, resulting in vowel reduction, hence [mɛ́inlənd]. Conversely, lack of secondary stress and concomitant vowel reduction is a sign that a semantic hierarchy has determined the particular phonetic form of the word, resulting in the full vowel [a] instead of a schwa. This is an immanent structural fact of the variety of English of those speakers who implement this pronunciation. When the constituent in question retains a secondary stress, and the word is consequently pronounced [mɛ́inlànd], the semantic hierarchy accords (nearly) equal rank to both constituents. Its status as a compound is less distinct than in the form with no secondary stress and a schwa. Speakers whose idiolect manifests the unreduced form of the vowel, therefore, have a different understanding of the hierarchical status of the two constituents in question from those whose speech here manifests vowel reduction.

MICHAEL SHAPIRO