have you found a plausible way to explain something unusual?
have you found a plausible way to explain something unusual?
in my conlang, bayerth, there are zero irregular verbs (irregularity is limited to one thing, inflection of pronouns); even 'to be' is regular; but I found a thing that justifies this to a degree; I split to be up amongst many different verbs; creating different words for over a dozen senses of the English word; the idea is that they then split up the frequency of use that a copula atracts amongst themselves; limiting how often any one is used; which helps check the tendency to irregularity; have any of you found ways to have a naturalistic justification for something unusual like how bayerth handles "to be"?
Re: have you found a plausible way to explain something unusual?
...but then aren't you just 'explaining' something perfectly ordinary and not really in need of explanation (not having irregular verbs) by means of something utterly bizarre and implausible and requiring some sort of explanation (the speakers being pedantic philosopher-kings who are able to consistently distinguish twelve different meanings of 'is')?
Maybe we should step back and ask why 'to be' is sometimes an irregular verb:
1. It's not a verb
That is, in some languages the copula is not treated as a verb; it may have totally different morphology and/or syntax from verbs. It may derive from a totally different part of speech, like a demonstrative, for example, or a pragmatic particle. Likewise, existence is semantically distinct from action (as anti-Meinongian philosopher-logicians like to chant: "existence is not a predicate!"), and existential verbs may not be verbs, or may be weird verbs.
However, this source of irregularity can easily be avoided by making your copula and/or existential an ordinary verb.
2. They're often defective
Due both to semantics (the philosophy of 'used to be an X' vs 'is a former X' is complex, for instance) and diachronics (often not originating as verbs), 'be' verbs are often lacking parts of their paradigms (other tenses, other moods, etc). This is itself an irregularity, and can in turn provoke more irregularities by causing either novel morphological reconstruction of the missing parts, or suppletion from other sources. Famously, the infinitive, participles and past tense of English 'to be' are provided by suppletion.
However, this source of irregularity can easily be avoided by making your copula and/or existential be non-defective from the start, or seemlessly repairing the defects (which is actually easier if verbs in general are more regular, and hence easier to analogise from).
3. They semantically blend into other verbs
"To be" has several senses; it's semantically weird; it's sometimes irregular; and it's so common that it often lacks rhetorical force. It's therefore often reinforced by, replaced by, or merged with other verbs with related meanings. If multiple mergers occur incompletely, you end up with a highly suppletive verb. So, the English (and German) verb 'to be' represents three PIE verbs - one is the original 'to be', one is a verb originally meaning 'to become', and one is a verb originally meaning 'to stay'. Semantically, "you became X", "you stayed X" and "you were X" are all very close together! Similarly, Spanish and Irish have given some senses of "to be" to a verb that originally meant "to stand" (Old English likewise had a semantic split in the present tense, but this has been lost).
However, this source of irregularity can easily be avoided by making your copula and/or existential not blend into other verbs. Or, any irregularity can be swiftly eliminated by completely replacing it by another verb.
4. They can be old verbs
Not withstanding the above, it's possible for "be" verbs to survive a long time. As a result, they may lack novel mechanisms found in newer verbs; in some cases, either all other old verbs may have been replaced by new verbs, or the novel mechanisms may have spread to all other verbs. In the case of English, the -m ending of the first person singular of 'to be' is the last vestige of the old athematic verbal inflection of PIE; all other verbs either come from the newer thematic paradigm or else have analogised the endings of that paradigm.
However, this source of irregularity can easily be avoided by making your copula and/or existential not be old verbs, or by not having a major shift in verbal morphology at any point.
5. They are often semantically low-salience, and hence discourage excess marking
One reason 'be' verbs may not pick up features that spread to other verbs is that their forms are often semantically of little importance compared to other verbs, and hence there's an incentive not to burden them with extraneous markings, even when those markings have become obligatory, or even meaningless, in other verbs.
However, this source of irregularity can easily be avoided by making your copula and/or existential share the same markings as other verbs by analogy, or by not introducing novel markings to other verbs in the first place.
6. They are often phonologically weak
Because they are low-salience, 'be' verbs are often destressed, and may be cliticised. These things can result in irregular sound changes, or simply regular sound changes that only occur in unstressed words, or word-internally - for instance, the 't' of 'ist', which usually preceded a consonant form, creating an stC cluster, has been lost in English, even though truly word final -st clusters don't lose the -t, and cluster reduction doesn't normally phonemically occur across word boundaries. [in German, the -t is still written, but is often not pronounced, presumably for similar reasons] When 'be' verbs ARE stressed, the result can be an artificial, reconstructive stressing of an underlying reduced form, which can result in weird and variable outcomes.
The lack of stress (and consequent reductio to schwa, with secondary reconstruction of a full vowel under the influence of surrounding consonants) is also probably why 'am' has an 'a' in it: it started with *i, but by Old English the forms 'eom', 'eam', 'am', 'aem' and 'iom' were all found, reducing to 'am' and 'em' by Middle English, and only 'am' in Modern English.
However, this source of irregularity can easily be avoided by making your copula and/or existential not be phonologically weak, or by not having weak words undergo different sound changes.
7. They're often short
Presumably because they're both phonologically weak (encouraging reduction) and semantically weak (discouraging long words), 'be' verbs are often short. If they're shorter than most verbs (eg if most verbs are bisyllabic but 'be' is a monosyllable), this can result in strange outcomes by changing where the stress falls, or by providing less material that can be safely removed without mergers.
However, this source of irregularity can easily be avoided by making your copula and/or existential similar in length to other verbs, or not having sound changes that work differently with shorter words.
7. Because they are high-frequency words, they can retain (or even gain) categories lost in other verbs
'Be' verbs can retain things like dual number or subjunctive moods even when other verbs lose them.
However, this source of irregularity can easily be avoided by making your copula and/or existential act like other verbs in this way.
8. Irregularities can arise through regular sound change, as in other words
Most irregularities arise through the application of regular sound changes: small changes between original verb forms can lead to substantial 'irregularities' later on (see: English irregular verbs, most of which derive from a very simple regular vowel alternation in PIE).
However, this source of irregularity can easily be avoided by just not having sound changes that do this.
9. Because they are high-frequency words, irregularities that do arise can survive longer than in low-frequency verbs
This is what you've presumably noticed, and why you want to reduce the frequency of your 'be' words (which I think would just encourage mergers and suppletions, increasing irregularity!).
But this is of course only a delaying factor, not a preventative one. Analogy still works to eliminate irregularities. For instance, in Old English the plural of 'am' was 'sind', phonologically irregular (though diachronically predictable*). It had survived longer than equivalent irregularities in other verbs, but it was swiftly eliminated in Middle English (by borrowing a form from Norse that looked more like it was part of the paradigm).
Likewise, you could eliminate irregularities in your conlang simply by applying analogy.
In short: the copula and existential are not inherently condemned to be irregular, and in many languages are not. There are a number of factors that can help to produce irregularity in them, but none of them need to apply in any given language. Likewise, they are more protected from the elimination of irregularity through analogy than most other verbs are - but 'more protected' is only relative, and there's nothing preventing any irregularities from being eliminated. Particularly as it's commonplace for other (and perhaps more regular) verbs to adopt these roles (probably even more so when the original inhabitants of these roles have become problematically irregular!).
*what probably happened, simplifying a lot, was:
3rd singular: *Hés-ti > Hésti > ésti > ist (> is)
3rd plural: Hes-énti > Hsénti > séndi > sind
Maybe we should step back and ask why 'to be' is sometimes an irregular verb:
1. It's not a verb
That is, in some languages the copula is not treated as a verb; it may have totally different morphology and/or syntax from verbs. It may derive from a totally different part of speech, like a demonstrative, for example, or a pragmatic particle. Likewise, existence is semantically distinct from action (as anti-Meinongian philosopher-logicians like to chant: "existence is not a predicate!"), and existential verbs may not be verbs, or may be weird verbs.
However, this source of irregularity can easily be avoided by making your copula and/or existential an ordinary verb.
2. They're often defective
Due both to semantics (the philosophy of 'used to be an X' vs 'is a former X' is complex, for instance) and diachronics (often not originating as verbs), 'be' verbs are often lacking parts of their paradigms (other tenses, other moods, etc). This is itself an irregularity, and can in turn provoke more irregularities by causing either novel morphological reconstruction of the missing parts, or suppletion from other sources. Famously, the infinitive, participles and past tense of English 'to be' are provided by suppletion.
However, this source of irregularity can easily be avoided by making your copula and/or existential be non-defective from the start, or seemlessly repairing the defects (which is actually easier if verbs in general are more regular, and hence easier to analogise from).
3. They semantically blend into other verbs
"To be" has several senses; it's semantically weird; it's sometimes irregular; and it's so common that it often lacks rhetorical force. It's therefore often reinforced by, replaced by, or merged with other verbs with related meanings. If multiple mergers occur incompletely, you end up with a highly suppletive verb. So, the English (and German) verb 'to be' represents three PIE verbs - one is the original 'to be', one is a verb originally meaning 'to become', and one is a verb originally meaning 'to stay'. Semantically, "you became X", "you stayed X" and "you were X" are all very close together! Similarly, Spanish and Irish have given some senses of "to be" to a verb that originally meant "to stand" (Old English likewise had a semantic split in the present tense, but this has been lost).
However, this source of irregularity can easily be avoided by making your copula and/or existential not blend into other verbs. Or, any irregularity can be swiftly eliminated by completely replacing it by another verb.
4. They can be old verbs
Not withstanding the above, it's possible for "be" verbs to survive a long time. As a result, they may lack novel mechanisms found in newer verbs; in some cases, either all other old verbs may have been replaced by new verbs, or the novel mechanisms may have spread to all other verbs. In the case of English, the -m ending of the first person singular of 'to be' is the last vestige of the old athematic verbal inflection of PIE; all other verbs either come from the newer thematic paradigm or else have analogised the endings of that paradigm.
However, this source of irregularity can easily be avoided by making your copula and/or existential not be old verbs, or by not having a major shift in verbal morphology at any point.
5. They are often semantically low-salience, and hence discourage excess marking
One reason 'be' verbs may not pick up features that spread to other verbs is that their forms are often semantically of little importance compared to other verbs, and hence there's an incentive not to burden them with extraneous markings, even when those markings have become obligatory, or even meaningless, in other verbs.
However, this source of irregularity can easily be avoided by making your copula and/or existential share the same markings as other verbs by analogy, or by not introducing novel markings to other verbs in the first place.
6. They are often phonologically weak
Because they are low-salience, 'be' verbs are often destressed, and may be cliticised. These things can result in irregular sound changes, or simply regular sound changes that only occur in unstressed words, or word-internally - for instance, the 't' of 'ist', which usually preceded a consonant form, creating an stC cluster, has been lost in English, even though truly word final -st clusters don't lose the -t, and cluster reduction doesn't normally phonemically occur across word boundaries. [in German, the -t is still written, but is often not pronounced, presumably for similar reasons] When 'be' verbs ARE stressed, the result can be an artificial, reconstructive stressing of an underlying reduced form, which can result in weird and variable outcomes.
The lack of stress (and consequent reductio to schwa, with secondary reconstruction of a full vowel under the influence of surrounding consonants) is also probably why 'am' has an 'a' in it: it started with *i, but by Old English the forms 'eom', 'eam', 'am', 'aem' and 'iom' were all found, reducing to 'am' and 'em' by Middle English, and only 'am' in Modern English.
However, this source of irregularity can easily be avoided by making your copula and/or existential not be phonologically weak, or by not having weak words undergo different sound changes.
7. They're often short
Presumably because they're both phonologically weak (encouraging reduction) and semantically weak (discouraging long words), 'be' verbs are often short. If they're shorter than most verbs (eg if most verbs are bisyllabic but 'be' is a monosyllable), this can result in strange outcomes by changing where the stress falls, or by providing less material that can be safely removed without mergers.
However, this source of irregularity can easily be avoided by making your copula and/or existential similar in length to other verbs, or not having sound changes that work differently with shorter words.
7. Because they are high-frequency words, they can retain (or even gain) categories lost in other verbs
'Be' verbs can retain things like dual number or subjunctive moods even when other verbs lose them.
However, this source of irregularity can easily be avoided by making your copula and/or existential act like other verbs in this way.
8. Irregularities can arise through regular sound change, as in other words
Most irregularities arise through the application of regular sound changes: small changes between original verb forms can lead to substantial 'irregularities' later on (see: English irregular verbs, most of which derive from a very simple regular vowel alternation in PIE).
However, this source of irregularity can easily be avoided by just not having sound changes that do this.
9. Because they are high-frequency words, irregularities that do arise can survive longer than in low-frequency verbs
This is what you've presumably noticed, and why you want to reduce the frequency of your 'be' words (which I think would just encourage mergers and suppletions, increasing irregularity!).
But this is of course only a delaying factor, not a preventative one. Analogy still works to eliminate irregularities. For instance, in Old English the plural of 'am' was 'sind', phonologically irregular (though diachronically predictable*). It had survived longer than equivalent irregularities in other verbs, but it was swiftly eliminated in Middle English (by borrowing a form from Norse that looked more like it was part of the paradigm).
Likewise, you could eliminate irregularities in your conlang simply by applying analogy.
In short: the copula and existential are not inherently condemned to be irregular, and in many languages are not. There are a number of factors that can help to produce irregularity in them, but none of them need to apply in any given language. Likewise, they are more protected from the elimination of irregularity through analogy than most other verbs are - but 'more protected' is only relative, and there's nothing preventing any irregularities from being eliminated. Particularly as it's commonplace for other (and perhaps more regular) verbs to adopt these roles (probably even more so when the original inhabitants of these roles have become problematically irregular!).
*what probably happened, simplifying a lot, was:
3rd singular: *Hés-ti > Hésti > ésti > ist (> is)
3rd plural: Hes-énti > Hsénti > séndi > sind
Re: have you found a plausible way to explain something unusual?
I don't think I ever do really unusual things, so I've never been asked or felt the need to explain something diachronically.
That said, maybe my standards for something to qualify as "unusual" are quite high?
I once had a conlang where
farala 'to fly' (stressed on fa-) vs. farāla 'flight' (stressed on -rā-)
> fárala vs. farála
timelo 'to freeze' (stressed on ti-) vs. timēlo 'frozen, act of freezing sth' (stressed on -mē-)
> tímelo timélo
Or perhaps the loss of a word-final consonant, which could pull stress to the last syllable:
gipainya 'to steal; stolen; a theft' (stressed on -pai-) vs. gipainyak 'a thief' (stressed on -nyak)
> gipáinya vs. gipainyá
Enough of these instances and analogy got to apply even in words where verbs should end up with a non-predictable stress:
irāme 'to laugh; laughed at; laughter' (stressed on -rā-)
> írame 'to laugh' vs. iráme 'laughed at; laughter' (effectively splitting into two)
That said, maybe my standards for something to qualify as "unusual" are quite high?
I once had a conlang where
Does this qualify as unusual, even though it's kind of attested in Ancient Greek? Maybe it does for some people. I didn't feel the need to come up with diachronics to justify it, but at any rate, in my mind this situation came about after the loss of vowel length that previously helped predict the location of stress in nouns (nouns could be derived by lengthening the second-to-last vowel, within the root, besides the use of prefixes):I had verbs hold regressive stress (as far back as possible depending on syllable structure, often on the third-to-last syllable), while nouns would have phonemic stress (often on the second-to-last syllable, sometimes the last one). This meant there were many pairs of a verb and a related noun that differed only in the location of stress: fárala 'to fly', farála 'flight', gipáinya 'to steal; a theft', gipainyá 'a thief'.
(I was later informed that Ancient Greek has something similar, with a pitch accent whose location in a word is largely regressive and predictable in verbs, but unpredictable in nouns and basic adjectives.)
farala 'to fly' (stressed on fa-) vs. farāla 'flight' (stressed on -rā-)
> fárala vs. farála
timelo 'to freeze' (stressed on ti-) vs. timēlo 'frozen, act of freezing sth' (stressed on -mē-)
> tímelo timélo
Or perhaps the loss of a word-final consonant, which could pull stress to the last syllable:
gipainya 'to steal; stolen; a theft' (stressed on -pai-) vs. gipainyak 'a thief' (stressed on -nyak)
> gipáinya vs. gipainyá
Enough of these instances and analogy got to apply even in words where verbs should end up with a non-predictable stress:
irāme 'to laugh; laughed at; laughter' (stressed on -rā-)
> írame 'to laugh' vs. iráme 'laughed at; laughter' (effectively splitting into two)
hīc sunt linguificēs. hēr bēoþ tungemakeras.
- Creyeditor
- MVP
- Posts: 5091
- Joined: 14 Aug 2012 19:32
Re: have you found a plausible way to explain something unusual?
I definitely see Sal's point, having a perfectly regular copula is probably not impossibly unnaturalistic. Still, I was wondering what the maximum number of 'be' verbs is in a given natlang. My conlang Kobardon has three unrelated roots for the locational copula, the nominal copula and the adjectival copula and I was never worried that these are too many. It's not a dozen but still, how many copulas can non-philosopher kings distinguish?
I often retroactively come up with explanations when I just messed up translations. In Kobardon this led to pre-nominal adjectives being non-restrictive and postnominal ones being restrictive.
I often retroactively come up with explanations when I just messed up translations. In Kobardon this led to pre-nominal adjectives being non-restrictive and postnominal ones being restrictive.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 2 3 4 4
Ook & Omlűt & Nautli languages & Sperenjas
Papuan languages, Morphophonology, Lexical Semantics
"Thoughts are free."
Produce, Analyze, Manipulate
1 2 3 4 4
Ook & Omlűt & Nautli languages & Sperenjas
Papuan languages, Morphophonology, Lexical Semantics
Re: have you found a plausible way to explain something unusual?
To answer the original question myself: mostly when there's a specific unusual things I want to do, attempt to explain them but never really succeed. However, what's much more common for me is to think of the explanation first, and then the unusual thing. That is, I think "oh, that would produce some unusual effects!" and then think what those effects could be, rather than starting with the unusual thing and having to find the explanation. [although sometimes I start with a general idea of where I want there to be something unusual, and only later fill it with a specific suggestion. For instance, in a language I've been thinking about recently, I knew I wanted some very strange things in some verb forms, but didn't have specific ideas for what. Then I had the idea of having some verbal constructions derive from reanalysed relative clauses, and this automatically provided several potential oddities to play with]
Yes, I'd say that that's unusual, and warranting an explanation. It's not bizarre - it's totally naturalistic* - but it's something that doesn't happen very often, I don't think.
I once had a conlang where [...]
Does this qualify as unusual, even though it's kind of attested in Ancient Greek?
*iirc some theorists say that it's impossible for different word classes to have different stress rules. The problem with this theory is that it's obviously bollocks, but that doesn't stop linguists, so...
Re: have you found a plausible way to explain something unusual?
I don't know, that wasn't a definite law being offered on my part, just doubt. And of course ultimately it's undefinable, because countless verbs could be found in at least one translation of at least one English sentence using 'be'. [eg it's common for more substantial verbs to be usable in the role of copulae describing states - "the river is cold" > "the river runs cold", "the tree is tall" > "the tree stands tall", "the fire is hot" > "the fire burns hot", etc. Are 'run', 'stand' and 'burn' all "verbs for 'to be'?" If not, why not? I could certainly imagine a language* in which most or all adjectival predication used verbs of this kind, which would take on a sort of classificatory function (trees stand, rivers run, fires burn, etc). But I wouldn't really count this as having a vast number of verbs for 'to be'].Creyeditor wrote: ↑29 Jul 2022 23:26 I definitely see Sal's point, having a perfectly regular copula is probably not impossibly unnaturalistic. Still, I was wondering what the maximum number of 'be' verbs is in a given natlang. My conlang Kobardon has three unrelated roots for the locational copula, the nominal copula and the adjectival copula and I was never worried that these are too many. It's not a dozen but still, how many copulas can non-philosopher kings distinguish?
I was thinking more of cases where there are multiple verbs the primary function of which is to serve in roles translated by 'be' in English, and where the choice of verb reflects a specific change in meaning in the relation, rather than a classificatory selection by subject or predicate.
In that case.... I don't know? Two is obviously common. I think three would be completely believable (eg copula vs existential vs locational). Four doesn't seem too difficult (split the copula into nominal and adjectival predication). One or more of the copular verbs could be split by aspect (eg permanent vs temporary, or temporary vs durative). You could also split the nominal copula into classification vs identification. You could perhaps distinguish the qualificatory existential from the pseudo-predicative (that is, "there is such a thing as a horse" vs "this horse exists")? So yes, you could get up to twelve distinctions even without getting into weird marginal functions of the English verb...
...I'm just skeptical that ALL those distinctions would be made at the same time in the same language, using a distinct verb for each one (rather than repurposing one distinction to other purposes - eg the Irish split is usually explained as nominal predication vs adjectival and prepositional predication, but it's also sometimes used to encode an essential vs accidental dimension, AIUI).
Re: have you found a plausible way to explain something unusual?
Yes, pronouns are a way that people can explain something unusual.
Anyway, I have no examples except that I plan on giving dense meanings to certain roots, like "to remember (usually after a long absence) knowing something, all the while being flushed with a graphic fullness of the last corresponding event."
Anyway, I have no examples except that I plan on giving dense meanings to certain roots, like "to remember (usually after a long absence) knowing something, all the while being flushed with a graphic fullness of the last corresponding event."
𖥑𖧨𖣫𖦺𖣦𖢋𖤼𖥃𖣔𖣋𖢅𖡹𖡨𖡶𖡦𖡧𖡚𖠨