In a recent article in ET 109 Michael Bulley (2012: 35) presents ‘the permutations for monosyllables in common use in standard British English having the phonetic pattern: single consonant + short vowel + single consonant’, using the OED as his source for data. While acknowledging the usefulness of Bulley's tables for pedagogical purposes, I wish to offer a very distinct answer to the question why certain words are missing from the contemporary English lexicon although they could have occurred as possible words.

There are three reasons, and they have nothing to do with ‘phonetic overload’ or aesthetic aspects. The first scenario is that the ancestors of English had no such patterns in their lexicon. This is the reason why English words do not begin with /ŋ/. One special case of this scenario is when English simply lost words or certain combinatorial possibilities along the way, so we no longer have a given pattern because the words have died out. The English spoken in the tenth century allowed combinations that are no longer possible today: wlaco /wl/ ‘tepid, cool’ was well-formed. Similarly, although words like knife and know have remained, the initial combination [kn] has died out. Another special case of this first scenario is that English came up with combinations or sounds that did not previously exist in Germanic, such as the /ʧ/ in chip or the vowel in moan. The second possible scenario is that English has not had extensive contact with other languages from which it could borrow such words. This is a subtle possibility. One can say that English does not have mip or sov because the languages from which it could borrow them also lack these words (or if they do, English speakers did not feel the need to borrow them). The third possible reason is that native speakers of English have not come up with such words because they apparently fare well with the millions of words they already have. People in the marketing and branding business have not yet come up with products they want to call sov or mip. It is not very productive to meditate on reasons why they haven't borrowed or created them yet since these processes are unpredictable. We'll know that the words exist when they first appear.

More importantly, however, it is not phonetic overload, aesthetic factors or English ears preferring certain sounds while rejecting others that account for these missing but possible words, as the author speculates (2012: 36). This is controversial terrain. Bulley evokes phonetic overload for the lack of *pag because there is pack–bag–back and asks: ‘Could we cope with a pag as well?’ Yes, we could. Its absence has nothing to do with phonetic overload since such four-member sets do occur in English (bug–buck–pug–puck etc) and it does not depend on the set of words pack–bag–back. How is this set a better reason for the absence of *pag than a set like tag–tack–pack would be? As for aesthetic factors, he speculates (2012: 36) that /ʧ_ʧ/ words with a short vowel are absent because the ‘quick repetition’ of this sound hurts ‘the English ear’. Note, however, that there are other ‘quick repetitions’ in English (dad, tit, cock, judge, nun etc). How far should the two /ʧ/ sounds be from each other to be far enough to allow for repetition? The only reason why the author does not consider the word church /ʧɜ:ʧ/ is because it has a long vowel /ɜ:/ excluded from the article.

As a final remark, it is very revealing to calculate how many such gaps there are in the first place. Approximately 680 words are included in Bulley's tables. If we take into consideration all the 24 consonants of English and let them loose in all positions, the rough product of the consonants runs to 24 × 6 vowels × 24 = 3456. Many possibilities are ruled out by various phonotactic constraints such as ‘final /h j w r/ is not allowed in British English’ and ‘/ŋ/ is only possible finally’. Discounting these, we still get 2760 possible English words of the CVC type. There are thus surprisingly many possible but missing words – all up for grabs. Or could that be graps?