Strategies for dealing with limited/simple phonologies

Question

What problems arise when creating a language with an extremely simple phonology and what are some good strategies for dealing with them?

I'm calling this type of phonology extremely simple for the purposes of this question, but limited or rigid might be a better fit. I'm also not sure whether "phonology" is the right term here or if something like "phonological inventory" would be better. Using "phonemes" or "phonemic inventory" is close, but does not cover phonotactics and suprasegmental features.

---

An extremely simple phonology is one that does not allow the speaker to pack many bits of information in each syllable/mora/unit of speech time. Prototypical extremely simple phonologies have some or all of the following features:

• Few segmental contrasts
• Restrictive phonotactics
• Limited or no phonemic use of gemination or vowel length
• Limited or no phonemic use of tone

Phrased another way, a simple phonology in the sense of the question is what you would get if you started with an inventory like Hawaiian's, Rotokas', or Pirahã's and removed contrastive vowel length or contrastive tone.

Conlangs like Toki Pona have extremely simple phonologies.

What are some common problems that appear in languages with phonologies like this? What are some strategies for dealing with them? One example of a possible "problem" might be a large number of homophones.

Answer 1

One major problem is that a small phoneme inventory leads to longer words, as you have fewer short ones available. This kind of relates to the point you already mention, namely homophones: here you simply re-use a word form for multiple meanings.

Words would also sound more alike, even if different. This is one of the main problems I have with Hawai'ian (of course the vocab is mostly unrelated to the Western European languages I know), that words are hard to recognise for a learner, and then it's even harder if individual phonemes are different and completely change the meaning of the word.

This is fine for toki pona, which only has a small inventory of words anyway, but any more extensive language would quickly run into longer words or phrases.

Answer 2

I think Oliver pretty much nailed it on the head. There are a couple of ways around the problems that were mentioned. The first way is to have a highly inflectional language; this would give an explanation for why words are similar, but the words would get longer. Because you would already be expecting to have longer words, this could at least be turned into a feature rather than a detriment. The second way is to allow for less 'traditional', let's say, syllable structures. Think of the English word 'strengths' (/stɹeŋθs/) which has a CCCVCCC structure. Using a much more restrictive (C)V(N) will only make the generations of new words more difficult.

Another problem is the potential overlap of meaning. The word moku in toki pona basically means to Ingest but it also means something that is ingestible (to Eat, to Drink, Food, Beverage). If you don't mind having words having a number of meanings which are generally related, then this shouldn't be a problem. If you want a free couple of words, don't put 'to Be', 'of/from', or (more extremely) 'to Have' into the language. That will save you 3 words right there for different concepts.

Answer 3

Disclaimer: I'm not trying to give a complete answer, just adding on.

I've encountered this exact problem with a project I'm working on. If the number of possible sounds per syllable is limited, then limiting syllables is going to be one of the main challenges. Some approaches to limit syllables are:

• Assign a separate meaning to every possible syllable that could be pronounced. It's best if the meanings are as distant as possible from each other.
• Avoid synonyms, similar to the previous point, where multiple different sounds map to the same meaning.
• Allow for things like optional contractions, that shorten the spelling or pronunciation.
• Use strict word order rules, as opposed to grammatical particles/prefixes/suffixes, when possible. If a subject is always at the beginning of a sentence, it won't need to be marked as a subject, for example.
• When deriving new words by combining existing ones, use very small fragments of root words as opposed to the whole thing. For example, instead of sand + castle --> sandcastle, do sand + castle --> sandle/sastle.
• In addition to assigning meaning to syllables, an extreme approach would be to go down to the letter level, approximate meanings for each letter. This can help in the compound word scenario with picking important letters out of a root as opposed to the whole word.

Answer 4

Elision would become the norm.

Speakers of a language with a small inventory and strict CV phonotactics would speak rapidly to compensate for longer words, like speakers of Spanish or Toki Pona. They would elide weaker phonemes, and listeners would identify the dropped sound through their effect on neighboring sounds.

Start with the inventory of Rotokas and simplify it even further. Give it six consonant phonemes p, b~v, t~s, ɾ~d~z, k, ɣ~ʝ and three vowel phonemes a, i, u without contrastive length. You end up with eighteen syllables, some showing anticipatory assimilation before /i/: pa, pi, pu, va, vi, vu, ta, si, tu, da~ra, zi~ri, du~ru, ka, ci, ku, ɣa, ʝi, ɣu. This allows up to 5,800 three-syllable sequences (or more realistically a fraction of that).

In rapid speech, /i/ and /u/ might tend to drop out in syllables before or after a syllable that has /a/. This sort of elision is productive in Japanese and occurred diachronically with the short "yer" vowels in Russian, which had been /ĭ/ and /ŭ/ and were reduced to traces on consonants. Likewise, diphthongs may arise at the surface when /ɣ/ weakens. Slow, careful speech would bring them back, at least until reanalysis a century or two later causes the next sound shift.

To illustrate, populate the lexicon with a few English loans:

• /vurara/ "brother", [vɾaɾa]
• /titita/ "sister", [sista]
• /titaku/ "stack", [stak]
• /taki/ "touch", [tac]
• /raɣutu/ "route, doubt", [ɾaut]
• /vaɣutu/ "vote", [vaut]
• /patu/ "pat", [pat]
• /pati/ "pass", [pas]

Even though /patu/ and /pati/ have lost their final vowel, the listener can tell them apart because of how vowel /i/ has changed the consonant. Likewise with the final consonants in /titaku/ and /taki/.

Answer 5

It's spoken faster.

For natlangs, there's evidence that the amount of information conveyed per second is consistent across different languages. This is "information" in the sense of Shannon's information theory, measured in bits.

If you have a smaller phonology, then there are fewer possible "signals" (in Shannon's terms), so each signal conveys less information. Thus, you need more signals per second to convey the same amount of information.

The end result is:

• Words are longer, as mentioned by others
• More phonemes are spoken per second

This is one reason why English-speakers tend to perceive Japanese as being spoken very quickly. Japanese has a much smaller phonemic inventory than English (and a much more restrictive syllable structure), so the number of syllables spoken per second is dramatically higher. I imagine the same would apply to Hawai'ian and various other Polynesian languages, which are also famous for their small phonemic inventories, though I'm not aware of any studies on them specifically.