I think it was in J. D. Salinger’s Raise High the Roofbeam, Carpenter that the word jeechet first appeared in print. It has always fascinated me because it is clearly one single word phonologically (a phonological word is easily defined as a series of linguistic sounds bearing a single accent). However, this ‘word’ corresponds to an entire four-word sentence!
Now, before you say that this is not a word but just the result of lazy speakers slurring their speech, let me assure you that linguists can track every single change from the sentence to the word using common rules of English phonology, i.e. rules that occur widely elsewhere and throughout the language. Here they are for your amusement and edification.
The first rule is that function words (monosyllabic pronouns, auxiliaries, prepositions, etc.) like did are rarely accented except for emphasis which is irrelevant here. The second rule is that unaccented vowels occurring before the accented syllable like the [i] here are regularly dropped in English, e.g. p’lice for police, s’pose for suppose. Since did is unaccented, it attaches to [you] for accent and the [i] then disappears, giving us
ddyou eat yet
However, since English (unlike Italian, for instance) does not tolerate double consonants, [dd] regularly reduces to [d] resulting in
dyou eat yet
Since you is another function word, it isn’t accented either and is regularly reduced to yê where [ê] represents a schwa, pronounced, roughly, [uh] dyê eat yet. However, since it is not accented, it must attach to the following word for accent, giving us
The only accent in this sentence is on eat which means that the vowel [ê] is now an unaccented vowel preceding the accented one and so falls to the ax of the second rule mentioned above, resulting in
Next, the combination [dy] regularly reduces to [j] and [ty] to [ch], e.g. mature [mêtyur > [mêchur] and picture [piktyur] > [pikchur]. Since the accent is on eat in this sentence, both the [dj] and [ty] are subject to this rule, which reduces our sentence further to
pronounced [djeechet]. Of course, the sound [j] is a combination of [d] + [zh], the sound of the Z in azure. This makes the [d] redunant, giving us
One reason we can’t determine the number of words in a language is because a phonological word (the sound part) does not always directly correspond to a semantic word (the meaning). According to Dr. Language at yourDictionary.com (also me), “I would have” comprises 3 distinct sounds and meanings but “I’d've” is a single two-syllable phonological word that matches the same three meanings—one word or three?
Speaking a language involves a complex set of mental activities in different parts of the brain each of which follows its own rules. The output of these rules are plotted onto the input of others in ways linguistics is still exploring. One of the most remarkable aspects of language is the surprising variety of rules and interaction of rules that the brain must carry out in order for us to express ourselves and be understood.
No other ‘word’ in the English language exemplifies the labrynthine nature of the levels of grammatical rules and their interactions better than jeechet.