The breakdown of the word "strength" or "cheap" or "sheep"

Question

So it turns out that sometimes consonants in a sequence can be called single consonants (e.g. d͡z), or consonant "clusters". But the main reason for calling d͡z a single consonant is because it "functions as a single slot in a word".

In order to learn beyond the simple "any single character not a vowel is a consonant, and any unbroken sound is a syllable" mindset, I'm wondering how the following words can be broken down:

• strength
• cheap
• sheep
• cloud

For sheep, the ee isn't a long sound (I don't think), I don't think it counts as a diphthong. The sound sh is one sound in IPA /ʃ/. So really it is like ʃip, 3 "~sounds~", but 5 letters in English. And also 1 syllable. But I'm wondering if the concept of "slots in a word" come into play here in any way. For example, if this is a 3-sound-slot word, where "sh" is the first sound, "ee" is the next", and "p" is the last.

The word cheap is pretty much the same, but if you break apart "ch" into "tʃ", then it is tʃip, so 4 sounds, but with 3 sound slots, and 1 syllable. But really, tʃ is one "slot", so it would be better represented as t͡ʃ. I assume the weird spelling of ea vs. ee is just an English thing, not really too concerned about that atm, unless that would be called a diphthong.

Then strength is 7 letters but 1 syllable. But has 2 consonant clusters, str and ngth. Wonder what those are called, and what the timing information / structure is around how this word is said.

Finally, the word cloud has a diphthong I think, ou, since the mouth changes over two vowels to produce that sound. And it has 1 consonant cluster kl, so kl-ow-d. I'm not sure if there is terminology around the fact that, even though there are 2 sounds in the ou sound, it only takes up 1 position in the timing of the word. Also, since cloud has 3 "chunks", but only 1 syllable (and one of the chunks is a diphthong), not sure if there is terminology around that. Like, the syllable has a diphthong in it, as well as consonant clusters. And these clusters like kl could be considered one "slot", such as with d͡z, like k͡l. Just as str or ngth could be considered one "slot", like s͡t or ŋ͡ð. I don't see when to consider the consonants a single unit or slot, vs. separate consonants.

Basically trying to learn the terminology around this stuff a little better.

Answer 1

You have to start with well-grounded assumptions of what you are modeling. The most fundamental concept that you're dealing with is the segment. The claim is / has been that what we produce and hear in speech is some set of segments. These are, in fact, technically known as "phones". As far as "sheep" is concerned, you have to start with what's "given", namely actual sound. If you carefully look at / listen to a recording of the pronunciation of "sheep", you can identify portions of the waveform that correspond to particular sounds, so you find the "sh" part, the "p" part, and the "ee" part. Looking at the vowel part, you actually see that the vowel formants of the vowel change: F1 lowers and F2 rises. You can likewise look at "shape". The first thing you have to do then is convert the physical waveform into something discrete and symbolic, which is where a transcription comes in. Having gone through extensive training in phonetics, you determine that what is said is [ʃɪjp] and [ʃɛjp], and also "ship" is [ʃɪp].

You will also end up with sequences of phones like [pʰɪt], [spɪt], [pʰɪjk], [spɪjk], and eventually you want a more compact representation since it seems strange to sat that there are two kinds of p in English. An analysis of distribution may then suggest that there are rules which govern where particular phones can appear, since for example you can't say [pi], [pɪ], [pʰɪ] as a speaker of English (at least, not "speaking English" – anybody can learn to make sounds, with some training). From the set of phones, you can compress them down to a set of phonemes, which are the more basic sound that characterize a language.

The question you have to address in analyzing phones is, what is the best analysis of the word "shape". One theory is that this is the long vowel [e:]. Another is that it is the tense vowel [e] (these theories treat the vowel as a single thing). A third theory is that this is a diphthong -- two things, a vowel an another vocalic sound -- which could be represented as [ɛɪ, ei, ej, ɛj] and a few others. It is often difficult (sometimes impossible) to decide what system of phonemes underlies the individual phones of a language.

There has been a widespread belief that a given set of facts is consistent with only one grammatical analysis. There is also a widespread belief that form of a grammar is to some extent indeterminate, and even when the raw data is the same across speakers, some speakers may opt for an /ɛɪ/ analysis, others may opt for an /ei/ analysis. Analytic ambiguity is a mechanism of grammatical change. Credit to Jlawler for directing my attention to this fact.

Answer 2

What you're trying to do is break the words down into phonemes. This is different from the number of "distinct" sounds that appear on the surface!

strength /strɛŋθ/ [stʃɹɛŋkθ]
cheap /t͡ʃip/ [tʃip]
sheep /ʃip/ [ʃip]
cloud /klawd/ [kɫɑʊ̯ːd]

In some cases, there are sounds on the surface (like the [ʃ] and [k] in "strength") that don't correspond to any phoneme. In other cases, like button /bʌtən/ [bʌʔn̩], there are phonemes that don't correspond to any sound on the surface. The whole field of phonology is generally concerned with how exactly these underlying phonemes map to sounds on the surface.

The key to deciding when two "distinct" sounds are actually a single phoneme comes down to how they're distributed. In English, /t͡ʃ/ doesn't act like /t/ followed by /ʃ/, so it's useful to call it its own phoneme. But /kl/ always just acts like /k/ followed by /l/, so treating that combination as a single phoneme just makes the theory messier without explaining anything better.