Which parts-of-speech have the most roots?

A forum for discussing linguistics or just languages in general.
Post Reply
User avatar
eldin raigmore
korean
korean
Posts: 5600
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Which parts-of-speech have the most roots?

Post by eldin raigmore »

I asked three questions on “Quick Questions” and they turned out not to be so quick.
I hope the conversation isn’t over yet!
So I’m copying the relevant posts over to start a new thread.
Spoiler:
eldin raigmore wrote:
09 Jul 2020 21:42
1. Does nearly every natlang that has both noun-roots and adjective-roots distinct from each other, have significantly more noun-roots than adjective-roots?

2. Does nearly every natlang that has both noun-roots and verb-roots distinct from each other, have significantly more noun-roots than verb-roots?

3. Does nearly every natlang that has both adjective-roots and verb-roots distinct from each other, have significantly more adjective-roots than verb-roots?

..........

I ask because according to the lexicographers of the most recent OED, English is an example of “yes” for all three questions.
Spoiler:
.
Ser wrote:
10 Jul 2020 06:05
eldin raigmore wrote:
09 Jul 2020 21:42
1. Does nearly every natlang that has both noun-roots and adjective-roots distinct from each other, have significantly more noun-roots than adjective-roots?

2. Does nearly every natlang that has both noun-roots and verb-roots distinct from each other, have significantly more noun-roots than verb-roots?

3. Does nearly every natlang that has both adjective-roots and verb-roots distinct from each other, have significantly more adjective-roots than verb-roots?
How does one decide whether a root is a verb-root or a noun-root?

If I look at an Arabic root, say, k-t-b, and its basic derived words like kataba 'to write' (using the CaCaCa transfix that creates action verbs) and kitaab 'book' (using the CiCaaC transfix that creates nouns), how do I decide whether I should call it a "verb-root" or a "noun-root"? You could argue that 'book' might be less basic and be a result noun of kataba, but I'd reply that Arabic doesn't normally use CiCaaC to derive result nouns, in fact, more often than not this a plural transfix of a singular created with CaCC(-a) or CaCaC (sall-a 'basket' pl. silaal 'baskets', jabal 'mountain' pl. jibaal), but kitaab is a singular with plural kutub. How does one know that kitaab isn't a borrowing with kataba based on it, or that kataba wasn't created after kitaab?

Or, for another example, sˤ-b-ħ, which derives the basic words sˤabuħa 'to be or become beautiful' (using the CaCuCa transfix that creates stative verbs) and sˤubħ 'dawn' (using the CuCC transfix that creates nouns). Could sˤubħ have meant 'beautiful' at one point, and sˤabuħa derived from that? Or n-dʒ-m, which produces nadʒama 'to appear in sight, arise' and nadʒm 'star' (the latter using the CaCC transfix that creates nouns). Or b-z-q, with bazaqa 'to spit' and buzaaq 'saliva'.

Or, if I look at Classical Chinese, does 自, pronounced zì in Mandarin, count as an adverb-root (or pronoun-root) because it can mean 'for oneself' or 'by oneself', or a verb-root because it can mean 'to be from [a place]'?

Or, if we look at English in fact, is "urge" a verb-root or a noun-root?
Spoiler:
Creyeditor wrote:
10 Jul 2020 08:51
eldin raigmore wrote:
09 Jul 2020 21:42
1. Does nearly every natlang that has both noun-roots and adjective-roots distinct from each other, have significantly more noun-roots than adjective-roots?

2. Does nearly every natlang that has both noun-roots and verb-roots distinct from each other, have significantly more noun-roots than verb-roots?

3. Does nearly every natlang that has both adjective-roots and verb-roots distinct from each other, have significantly more adjective-roots than verb-roots?

..........

I ask because according to the lexicographers of the most recent OED, English is an example of “yes” for all three questions.
Just from anecdotal experience, I would say that most these languages have more verbs than adjectives or nouns and less adjectives than verbs or nouns. Of course, you have to exclude proper names from the noun category, because otherwise, in some languages, you might just be able to add nouns over and over. And it's hard to draw the line. Place names? Names of Deities? Names of Planets? Etc, etc.
English might just be strange.
Spoiler:
Salmoneus wrote:
10 Jul 2020 20:54
Ser wrote:
10 Jul 2020 06:05
eldin raigmore wrote:
09 Jul 2020 21:42
1. Does nearly every natlang that has both noun-roots and adjective-roots distinct from each other, have significantly more noun-roots than adjective-roots?

2. Does nearly every natlang that has both noun-roots and verb-roots distinct from each other, have significantly more noun-roots than verb-roots?

3. Does nearly every natlang that has both adjective-roots and verb-roots distinct from each other, have significantly more adjective-roots than verb-roots?
How does one decide whether a root is a verb-root or a noun-root?
A good question, of course.

My interpretation of the question, though, is just that a noun root produces nouns - by a productive and semantically predictable non-compound process* - and a verb root produces verbs, and so forth. A language then 'has both noun-roots and adjective roots distinct from each other' when there are some noun roots that are not adjective roots, and/or vice versa.

I assume eldin does not intend to rule out all languages where some noun roots are also (or are homophonous with) verb roots, etc - since that would probably be all languages. "Urge" in English is in this way both a noun root and a verb root (and arguably an adjective root); however, 'octopus' is just a noun root, and 'explode' is just a verb root.

[a bigger problem in English is that all nouns can be verbs and vice versa: it would be perfectly possible for people to start writing "she octopused through the narrow gap in the fence" or "our media research reveals 42% more significant explodes in the last quarter of activity". In this sense English could be argued to be a noncategorial language. However, I think it's still meaningful to talk of distinct noun and verb categories in English, because there are some words that are not used as verbs, and some that are not used as nouns, and if they come to be used in those ways it will not be wholly predictable what their semantics will be; thus I think in English it makes more sense to talk of noun and verb categories with unrestricted zero-derivation, rather than noncategorialness]

*by 'non-compound process', I mean a process to produce, say, a verb from a root that can NOT be broken down into, say, a noun-forming process and then a distinct noun-to-verb-deriving process.


Regarding your Arabic examples, I would also distinguish between an etymological or lexicographical 'root' - useful in grouping words in dictionaries - and a genuine morphosyntactic root. I think X is only the morphosyntactic root of Y if Y can productively and predictably be derived from X; merely looking similar raises, as you point out, questions about etymology that should not be relevant to the synchronic analysis. So from what you say, it sounds like, for example, the s-b-h 'root' is not actually the morphosyntactic root of the word for 'dawn', even if it is the etymological root, because the derivation is not predictable (i.e. semantically transparent).


-------------


Going back to eldin's questions:

'3' is absolutely NOT the case, at all. There are many languages with distinct but very small true adjective classes (many Austronesian languages, for a start).

I believe '2' isn't true either. AIUI, in many North American polysynthetic languages the translations of many English nouns are in fact analysable as inflected verbs, and distributionally indistinguishable from deranked verbs (so, to invent an example, verbal "it is slithering" where English has nominal "snake"; "I saw the snake" would then be translated by something that would be identical in form with the translation of "I saw the one that was slithering"). I can't guarantee that this is true, but it's what I believe I've heard. AIUI, these languages do indeed typically have a distinct noun class not analysable in this way, but this class is relatively small, and much smaller than the class of verbs.

And of course, if 2 is untrue, then 1 must be untrue also, if some of these languages fail to distinguish verb and adjective classes, which I suspect they do...
Spoiler:
eldin raigmore wrote:
11 Jul 2020 09:43
“Ser” wrote: How does one decide whether a root is a verb-root or a noun-root?
If it isn’t clear, that’s not one of the languages I’m asking about in that question. (Question 2).
“Creyeditor” wrote: Just from anecdotal experience, I would say that most these languages have more verbs than adjectives or nouns and less adjectives than verbs or nouns. Of course, you have to exclude proper names from the noun category, because otherwise, in some languages, you might just be able to add nouns over and over. And it's hard to draw the line. Place names? Names of Deities? Names of Planets? Etc, etc.
English might just be strange.
(Thanks for the most relevant and responsive of the answers so far!)
Of course only common nouns are meant; proper nouns don’t usually belong to paradigms that have roots IIANM.
But maybe I am mistaken and those two (questions 1 and 2) of my three questions are not well-posed;
in which case I’d need help to re-pose them.

I am hoping for an answer that’s not just anecdotal, but since I have at most one anecdote, anyone else’s anecdotes are still information I didn’t have before!

(Also; maybe English IS just strange!)
“Salmoneus” wrote: My interpretation of the question, though, is just that a noun root produces nouns - by a productive and semantically predictable non-compound process* - and a verb root produces verbs, and so forth. A language then 'has both noun-roots and adjective roots distinct from each other' when there are some noun roots that are not adjective roots, and/or vice versa.
Right, mostly. Maybe entirely.
My notion was that a noun-root was already a noun before any inflection or derivation was applied to it; it could appear “in its surface form” or “at spell-out”, assuming I’m using either of those terms correctly, as a noun, as-is.
By that notion, some kind of derivational process could be applied to a noun-root and produce an adjective.
Gold—>golden or child—>childish or man—>manly, for instance.

Similarly I was thinking an adjective-root would already be an adjective without any inflection or derivation.
Verbs and nouns could be formed from an adjective-root, but they’d clearly be derived or inflected;
viz. red—>redden or red—>redness.

And there are examples too numerous to mention more than a small fraction of, of derivations or inflections applied to verbs, to produce nouns or adjectives. Infinitives, gerunds, participles (active and passive), gerundives, agent-nominalizations, place-nominaliztions, instrument-nominalizations, patient-nominalizations, time-nominalizations, action-or-event-nominalizations, etc.

Possibly my notions weren’t good enough to ask these questions about; and/or possibly yours are better. Or, possibly, they’re both good to ask about, but they make different questions and would result in different answers.

I assume eldin does not intend to rule out all languages where some noun roots are also (or are homophonous with) verb roots, etc - since that would probably be all languages. "Urge" in English is in this way both a noun root and a verb root (and arguably an adjective root); however, 'octopus' is just a noun root, and 'explode' is just a verb root.
I was not intending to rule out those particular languages, but I was intending to rule out those particular roots (at least for the question or questions for which that root would have caused confusion).
If most of a language’s adjective-roots are also noun-roots or homophonous with noun-roots; or most of its noun-roots are also adjective-roots or homophonous with adjective-roots; then that language is not one I intended to ask about in question 1.
Likewise if most of its verb-roots are also noun-roots or homophones with noun-roots, or most of its noun-roots are also verb-roots or homophones with verb-roots, I didn’t intend question 2 to be about that language.
And sim for adjectives and verbs and question 3.

Perhaps I should only exclude those languages where “most” could be changed to “all”, in the above exclusions?
Or instead, perhaps I should exclude only those for which the “or” after the comma could be changed to “and”? E.g. for question 2 exclude those for which both most noun-roots are verb-roots AND most verb-roots are noun-roots?

Also; the difference between how you think (or how you thought I thought) about noun-roots vs adjective-roots etc. is an important one for applying the above exclusions; it seems likely that different languages would be excluded under the notion I had but didn’t (until now) state, vs the notion you stated and I quoted.

[a bigger problem in English is that all nouns can be verbs and vice versa: it would be perfectly possible for people to start writing "she octopused through the narrow gap in the fence" or "our media research reveals 42% more significant explodes in the last quarter of activity". In this sense English could be argued to be a noncategorial language. However, I think it's still meaningful to talk of distinct noun and verb categories in English, because there are some words that are not used as verbs, and some that are not used as nouns, and if they come to be used in those ways it will not be wholly predictable what their semantics will be; thus I think in English it makes more sense to talk of noun and verb categories with unrestricted zero-derivation, rather than noncategorialness]

*by 'non-compound process', I mean a process to produce, say, a verb from a root that can NOT be broken down into, say, a noun-forming process and then a distinct noun-to-verb-deriving process.
Yeah. Zero-derivation might be a problem; I was just assuming that for languages for which zero-derivation is a recognized “thing”, linguisticians had some way to tell it was derivation, and not another use of the same word. (Another solution, sometimes adopted, is just to say that the two different parts-of-speech are two different words; they’re just homophones. I have no criteria for telling when that’s legit and when it’s cheating.)

'3' is absolutely NOT the case, at all. There are many languages with distinct but very small true adjective classes (many Austronesian languages, for a start)
I asked for “most such languages”. Not “all such languages”. Is it true of a significantly-large enough majority of languages for which the question makes sense, that it couldn’t just be chance that that’s the majority?

OTOH I’d very much like to hear about a few of those languages; and maybe hear what the class of true adjective-roots is for one (or a few? probably too much to ask for) such language(s).
I believe '2' isn't true either. AIUI, in many North American polysynthetic languages the translations of many English nouns are in fact analysable as inflected verbs, and distributionally indistinguishable from deranked verbs (so, to invent an example, verbal "it is slithering" where English has nominal "snake"; "I saw the snake" would then be translated by something that would be identical in form with the translation of "I saw the one that was slithering"). I can't guarantee that this is true, but it's what I believe I've heard. AIUI, these languages do indeed typically have a distinct noun class not analysable in this way, but this class is relatively small, and much smaller than the class of verbs
I specifically had in mind excluding some level of polysynthetic languages; I thought maybe polysynthetic IV, but I don’t have their level-numbers’ definitions memorized, so FAIK I also meant to exclude polysynthetic IIIs as well?
As I understand it some really really polysynthetic languages might not have noun-roots; they have noun-morphemes but these are never independent morphemes nor even root-morphemes. I could be wrong; and even if I’m right they may be a minority.
But anyway I meant to exclude them from questions 1 and 2.
And of course, if 2 is untrue, then 1 must be untrue also, if some of these languages fail to distinguish verb and adjective classes, which I suspect they do...
To the best of my knowledge, if question 2 fails to make sense for some language, then questions 1 and 3 also don’t make sense for that language. AIUI if a language doesn’t distinguish between noun-roots and verb-roots, it can’t (or at least doesn’t) distinguish between noun-roots and adjective-roots, nor between verb-roots and adjective-roots.

There are, as i inderstand it, languages which, in spite of distinguishing between nouns and verbs, don’t distinguish between noun-roots and verb-roots. I was asking question 2 about other languages; those that do distinguish between noun-roots and verb-roots.
There are, as I understand it, languages that distinguish between nouns and verbs, but neither distinguish between adjectives and nouns, nor between adjectives and verbs. Questions 1 and 3 are not about those languages.

You personally, @Salmoneus, may not have needed that clarified. Actually maybe nobody who has responded so far needed that clarified. But maybe a later responder will appreciate my clarifying my question.
“Ser and Sal wrote about 3Cons” wrote: ...
K-t-b and other roots in Arabic and other triconsonsntal-root languages have roots that never appear in surface utterances.
I think I once read that Akkadian construct-state nouns are an exception? Or at least some 3Cons does sometimes use a triconsonantal root without any vowel transfixed as binyan or wazan? (Did I spell either of those correctly?)
Anyway, if I define a noun-root as one that can be used as a noun in a surface utterance without any inflection or derivation, and a verb-root as one that can be used as a verb without an6 inflection or derivation, and so on, then such roots in such 3Cons natlangs wouldn’t be the roots I’m asking about; and, if the majority of the roots in such a language can’t be used in surface utterances without some kind of affixations or inflection or derivation, then that language probably isn’t one of the languages I’m asking about.
(I think Sal already got that but Ser hadn’t, at least not before Sal replied!)

I think a bigger obstruction might come from languages with verbnouns or masdars.
If I understand them!
In these languages, I take it, most or many verbs have a root that is a noun?

As for the Mandarin adverb-or-pronoun zi that Ser asked about, and the verb zi also asked about, I would bet that’s one example where they are clearly two different words that are just coincidentally homophones. Etymology would prove or disprove that. If anyone knows the etymology for the two meanings, they can tell us whether I’d win or lose that bet!

It’s a lot more difficult for me to just declare the “oneself” or “by oneself” meaning of zi to be a pronoun or an adverb. In this case my consciousness of my ignorance prevents me making a guess. (I’m also conscious of my ignorance concerning the “be or come from a location” meaning, but in that case I think my guess is probably right anyway, and don’t hesitate to state it, because I think my ignorance is irrelevant just this once!)
However I have not asked about adverb-roots and have not asked about pronoun-roots. If adverbs or pronouns are also adjectives or nouns, not necessarily respectively, then “zi” might be one of the roots I was talking about. But I don't know whether the majority of roots in Mandarin or other Chinese languages are so ambiguous? If they are then that (or those) language(s) are outside the scope of (at least one of) my question(s).

As for English’s “urge”, I think Salmoneus’s response says as much as I could say, except better.

..........

Obviously I hope for yet more answer, from either a brand-new responder or someone who has already responded.

I may need help improving the questions.
If anyone thinks they could provide such help I’ll welcome the attempt!
Spoiler:
eldin raigmore wrote:
11 Jul 2020 11:58
And BTW I just noticed my questions aren’t quick questions anymore.
They should head up a new thread.
But it’s 5:57 AM here now and I haven’t had any sleep.
I’ll try to do that in several hours.

User avatar
Ser
sinic
sinic
Posts: 267
Joined: 30 Jun 2012 06:13
Location: Canada

Re: Which parts-of-speech have the most roots?

Post by Ser »

Salmoneus wrote:
10 Jul 2020 20:54
My interpretation of the question, though, is just that a noun root produces nouns - by a productive and semantically predictable non-compound process* - and a verb root produces verbs, and so forth. A language then 'has both noun-roots and adjective roots distinct from each other' when there are some noun roots that are not adjective roots, and/or vice versa.
Okay, that sounds reasonable.
Salmoneus wrote:
10 Jul 2020 20:54
Regarding your Arabic examples, I would also distinguish between an etymological or lexicographical 'root' - useful in grouping words in dictionaries - and a genuine morphosyntactic root. I think X is only the morphosyntactic root of Y if Y can productively and predictably be derived from X; merely looking similar raises, as you point out, questions about etymology that should not be relevant to the synchronic analysis. So from what you say, it sounds like, for example, the s-b-h 'root' is not actually the morphosyntactic root of the word for 'dawn', even if it is the etymological root, because the derivation is not predictable (i.e. semantically transparent).
Yeah, I agree. This would mean that e.g., with the dictionary root ʃ-h-l, we'd probably need to identify an adjective-root of bodily property ʔaʃhal 'dark blue [eyes]', from which the colour ʃuhla 'dark blue colour [of eyes]' is derived (not wholly predictably as CuCCa doesn't guarantee a colour noun, but close enough?), and another adjective-root ʃahil 'nimble, swift', from which the verb ʃahhala 'to speed sth up' is derived (regularly using the CaC:aCa transfix that derives causative verbs).

Arabic is always fun to bring up when it comes to derivational morphology, since its heavy use of transfixes often makes the direction of derivation not very obvious. To build on the ʃahil > ʃahhala example, the CaC:aCa transfix also derives transitive verbs from basic nouns, regardless of what shape the noun may have: raʔs 'head (of the body); leader' > raʔʔasa 'to name sb as head/leader', sˤuura 'photo' > sˤawwara 'to photograph sth'.



eldin raigmore wrote:
11 Jul 2020 09:43
Of course only common nouns are meant; proper nouns don’t usually belong to paradigms that have roots IIANM.
But maybe I am mistaken and those two (questions 1 and 2) of my three questions are not well-posed;
in which case I’d need help to re-pose them.

I am hoping for an answer that’s not just anecdotal, but since I have at most one anecdote, anyone else’s anecdotes are still information I didn’t have before!

(Also; maybe English IS just strange!)
If we don't mind going by very informal and fairly unreliable intuition here, I... honestly have no clear impression, using Sal's definition. It's not at all obvious to me that English exhibits noun-roots > adjective-roots > verb-roots in number, nor Spanish, French or Latin, nor Mandarin.

Also, I don't know if Salmoneus' definition is the same that the OED team meant, and whether it'd be valid to compare other languages with it. See more right below...
eldin raigmore wrote:
11 Jul 2020 09:43
My notion was that a noun-root was already a noun before any inflection or derivation was applied to it; it could appear “in its surface form” or “at spell-out”, assuming I’m using either of those terms correctly, as a noun, as-is.
By that notion, some kind of derivational process could be applied to a noun-root and produce an adjective.
Gold—>golden or child—>childish or man—>manly, for instance.

Similarly I was thinking an adjective-root would already be an adjective without any inflection or derivation.
Verbs and nouns could be formed from an adjective-root, but they’d clearly be derived or inflected;
viz. red—>redden or red—>redness.
Maaan, under that definition of yours, you'd pretty much just consider languages that are low in inflectional morphology, like English and Mandarin and Hawaiian. Pretty much every language normally called polysynthetic is out (and so all of Algonquian, Eskimo-Aleut, Iroquoian, Athabaskan, Caddoan), and also most European IE languages, and Finno-Ugric, Afro-Asiatic, Bantu and Mayan.

To clarify what I mean, although languages like German, French and Czech have many noun-roots as you define them, they also have a lot of word roots that would go unaccounted for, making observations of these languages useless. In these languages, there are many roots that just never appear without at least some derivational suffix, way too many. Think of all the roots that only appear as feminine nouns in German, with -e. Also, nearly all (if not all) Czech neuter nouns end in a derivational suffix after the root, and you'd also be excising more than half of the feminine nouns. Spanish has no verbs at all without some kind of unpredictable thematic vowel (cant-a-r 'to sing', cf. cant-o 'singing (action noun)', cant-e 'flamenco singing'), and under your definition only a few several root-nouns like pan 'bread' and juez 'judge' count.)

Basically, thinking about roots that are also words like "gold" and "red" and "sing" seems okay in a language like English, but it breaks down with something like Spanish because Spanish has no inflection-less derivation-less verbs like English "sing", and hardly any such nouns or adjectives too. It is still a question of some interest for languages low in inflection, but it didn't originally sound like this was what you meant...

I think Creyeditor is using a definition that is not like yours too, but more like Salmoneus' or similar.
Or, possibly, they’re both good to ask about, but they make different questions and would result in different answers.
Yeah, I think so.
K-t-b and other roots in Arabic and other triconsonsntal-root languages have roots that never appear in surface utterances.
I think I once read that Akkadian construct-state nouns are an exception? Or at least some 3Cons does sometimes use a triconsonantal root without any vowel transfixed as binyan or wazan? (Did I spell either of those correctly?)
I have no idea about Akkadian. Standard Arabic and Hebrew don't have any such nouns though. And you spelled binyan correctly, but the other word is wazn, with Arabic plural /ʔawzaan/ (with a glottal stop at the beginning).
I think a bigger obstruction might come from languages with verbnouns or masdars.
If I understand them!
In these languages, I take it, most or many verbs have a root that is a noun?
Not Arabic or Hebrew. Pretty much all Arabic verbs have a verbal noun that is in some sense kind of part of the verbal inflection, even though the stem isn't predictable in the three more basic awzaan. In basic verbs (wazn faʕala), the stem is outright very unpredictable; in wazn faʕʕala, it is usually but not always taCCiiC(-a) with an unpredictable presence of feminine -a, and in wazn faaʕala it is unpredictably masculine CiCaaC or feminine muCaaCaCa.
As for the Mandarin adverb-or-pronoun zi that Ser asked about, and the verb zi also asked about, I would bet that’s one example where they are clearly two different words that are just coincidentally homophones. Etymology would prove or disprove that. If anyone knows the etymology for the two meanings, they can tell us whether I’d win or lose that bet!

It’s a lot more difficult for me to just declare the “oneself” or “by oneself” meaning of zi to be a pronoun or an adverb. In this case my consciousness of my ignorance prevents me making a guess. (I’m also conscious of my ignorance concerning the “be or come from a location” meaning, but in that case I think my guess is probably right anyway, and don’t hesitate to state it, because I think my ignorance is irrelevant just this once!)
Axel Schuessler in his etymological dictionary of Chinese says the verb may have possibly been derived from the adverb/pronoun, but if it was, this happened before the period of Old Chinese (i.e. pre-Classical Chinese), and at any rate it is uncertain that it was.
But I don't know whether the majority of roots in Mandarin or other Chinese languages are so ambiguous? If they are then that (or those) language(s) are outside the scope of (at least one of) my question(s).

As for English’s “urge”, I think Salmoneus’s response says as much as I could say, except better.
I think Sal was talking about making such a root count as both a noun-root and a verb-root though, for example, rather than leaving it outside the scope.
hīc sunt linguificēs. hēr bēoþ tungemakeras.

User avatar
eldin raigmore
korean
korean
Posts: 5600
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Which parts-of-speech have the most roots?

Post by eldin raigmore »

I did rather expect I was asking about a minority of the world’s languages; though I thought it might be a sizable minority instead of a small one, and definitely thought it would include English.

And you bring up an important point I neglected to consider;
the definition I/we need to use is whatever the latest OED’s lexicographers were using when they announced that;
* somewhat over half of English’s roots are noun-roots
* around a third of English’s roots are adjective-roots
* around a seventh of English’s roots are verb-roots.

I wonder if I (or you or any of us) can track that down? As far as I remember the publication wherein I read those fractions didn’t mention what they were using to decide which roots are noun-roots and which are adjective-roots and which are verb-roots.
And, if and when we do find it, can we be sure it can be made to apply to many of the world’s languages? Say, 30%? Or 15%? Or 7.5%?

Post Reply