0

I am working on a conlang and I came up with a question that I can't find a good answer too.

How does sound change effect suffixes, as whenever I work on conlangs with suffixes to mark different cases and such, I keep having the suffixes diverge a lot due to sound changes causing them to change in very different ways depending on the word leading to the loss of cases as it just becomes a ton of seemingly irregular forms without totally clear connection.

Like, random case,

Ban + ton --> Banton --> Bandon --> Banden --> Baden

Feton --> Feson --> Fesom

Where the conditions of the root causes changes on the suffix. This exact case isn't used in any of my attempts, but was just used as an example of the kind of thing that happens when I attempt this.

So how exactly does this work in actual languages that leads to clear cases and conjugations remaining and not just, wildly diverging over time and losing any semblance of a case structure.

Improve this question
9
  • 2
    What makes you think sound changes make clear systems in real languages? It works just as you describe there too: sound changes cause various suffixes (cases, verbal inflections, etc.) to change, merge, diverge or disappear entirely all the time. That’s why English doesn’t have cases anymore, for example. The thing that stops case (and similar) systems from completely breaking into a thousand unrecognisable pieces is analogy, whereby users know that a particular ending is supposed to look a particular way, so the regular sound change isn’t allowed to have its way with the given form.
    – Janus Bahs Jacquet
    Commented Mar 2, 2023 at 21:13
  • I meant it due to observation of languages having recognized systems. Where they have various forms of cases, but it is in a way where they are recognizable as a case, my issue is that each case in the proto-language over time just become seemingly unrelated terms. Where for instance, the word for human becomes around twenty different words instead of one word with different endings.
    – Zoey
    Commented Mar 2, 2023 at 21:33
  • 2
    And that’s precisely what analogy helps prevent when it’s allowed – which is usually, but not always. The Old Irish verb system is an excellent example of a case where sound changes have run their course with very little analogy to stay them, and the result is that the conjugation of even so-called ‘regular’ verbs is very tricky to predict; e.g., do·sluindi ‘he denies/rejects’ when negated ‘regularly’ becomes ní·díltai ‘he doesn’t deny/reject’. But by and large, things don’t get that far before analogy kicks in. Later in the history →
    – Janus Bahs Jacquet
    Commented Mar 2, 2023 at 23:06
  • → of Irish, the verb system was affected by analogy and reanalysis, so now it’s much more regular. For this verb, the absolute form do·sluindi was lost, and the conjunct form ·díltai was generalised, so now the stem of the verb is diúltaigh everywhere. The ending -igh is also analogical: it’s the most common and productive way to form verbs in Middle and Early Modern Irish, and in this case it was simply tacked on to the end of the stem to make it look more like a regular verb. Analogy. Just like when the English plural in -er was lost as a productive way to pluralise nouns. →
    – Janus Bahs Jacquet
    Commented Mar 2, 2023 at 23:11
  • 2
    The question is: how do suffixes affect sounds? and not: How does sound change effect suffixes, Since you are constructing a language, you can do whatever you like as long as you follow your own rules.
    – Lambie
    Commented Mar 3, 2023 at 17:53

1 Answer 1

Reset to default
1

In real language, having affixes means that inevitably there will develop phonological rules which change the pronunciation of roots and affixes. The only way to avoid that is to not have affixes, and that's not something that can be controlled in real languages.

First, it is inevitable that a given morpheme (word) will be produced differently depending on what comes before or after. This is basic coarticulation 101. Second, if morpheme distribution isn't random, certain sequences are likely to be more common, and may get conventionalized. Third, especially when separate words get mashed together for form bi-morphemic words (prefix+root or root+suffix), i.e. when historically separate "the pig" becomes a single word /ð+pig/, you are more likely to develop conventional phonological rules that lighten the perceptual and production load of sub-optimal phonetic forms (all phonetic forms are ultimately sub-optimal by some standard).

The problem that you seem to be alluding to is that sometimes, if there are not many differences marked with affixes and not a lot of roots that the affixes combine with, you may not be able to parse the roots and affixes and figure out the phonological rules where /ban+ton/ → [baden]. English has very few productive affixes. Bantu languages on the other hand have very many, and there is no problem with learning those systems because there is so much data that tells the child what the roots and affixes are. A given root can have millions of inflectionally-related forms, and you clearly cannot memorize all of the forms, so you either get rid of morphological differences, or you keep the phonological rules easy to learn.

Suppose you have a language with 500 roots and 20 affixes, but you only have 2,000 root+affix combinations, and in that set there are only 10 examples where the root+affix combination diverges on the surface from the simple morpheme concatenation. You can memorize those special cases; you can also decide not to memorize those combinations, so you get analogical leveling. The hardest solution to learn, in this case, might be "also learn the rules". If the rule is trivial like "t→s/V__V", perhaps that much can be learned. An alternative (known as analogical leveling) is that the child thinks "I must have heard that wrong, it should be [feton] and they don't have that rule anymore. Children don't faithfully reproduce what the adults around them say.

Improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.