Triconsonantal Root Systems

The best topics from Languages & Linguistics, kept on a permanent basis.
User avatar
Mecislau
Avisaru
Avisaru
Posts: 491
Joined: Thu Jul 24, 2003 2:40 pm
Location: Maryland
Contact:

Post by Mecislau »

Grath wrote:Hurray! The thread hath returned!
Wouldn't sukeret come from "sugar" (sukar, if memory serves me well) rather than from "flood"? That would make more sense to me... Compare diabetes in Dutch: suikerziekte "sugar disease".
Hmm... On second thought, I think you're right. Again, this is one of the problems of homophonous roots: S-K-R means both "flood" and "sugar". Although actually, both of these would make sense in describing diabetes, just that the u in "sukar" would explain why there's a u in "sukeret".
Grath wrote:Anyway... Coolness! So here are some random questions:
1. What difference in meaning is expressed in the difference in vowel patterns of qibuts and qvutsa?
Qibutz "kibbutz" and qvutzah "group, team" both come from the root Q-B-TZ "collect, gather". As for the exact meaning of the patterns, I don't know without further examples of each one. The patterns I already know mostly come just from listing out whatever words I know of with the same root. And I don't know of any list of vowel patterns that're out there.
Grath wrote:2. What root does mishpacha come from?
Hmm. Good question. The original root was clearly *SH-P-CH, but this doesn't exist as a verb anymore (if it ever did). Some dictionary searching found me shifchah "female slave" from the same root, as well as pach "snare" (possibly related, but I don't know - if it is, it clearly never adopted the she- prefix.

User avatar
Noriega
Sanci
Sanci
Posts: 48
Joined: Tue May 04, 2004 7:41 am
Location: Skaane

Post by Noriega »

Can Hebrew words undergo metathesis, so that the root letters change order when applied to certain patterns? I guess that would somehow contradict the entire principle of roots (in the sense that the radicals are always in a certain order), but I guess that sometimes, people rather facilitate the pronunciation than look at the etymology.
Perhaps eventually all languages will evolve so that they include some clicks among their consonants – Peter Ladefoged

Jahai: /kpotkpɛt/ ‘the feeling of waking up to the sound of munching’

User avatar
Mecislau
Avisaru
Avisaru
Posts: 491
Joined: Thu Jul 24, 2003 2:40 pm
Location: Maryland
Contact:

Post by Mecislau »

Noriega wrote:Can Hebrew words undergo metathesis, so that the root letters change order when applied to certain patterns? I guess that would somehow contradict the entire principle of roots (in the sense that the radicals are always in a certain order), but I guess that sometimes, people rather facilitate the pronunciation than look at the etymology.
No, not really. Not without the entire root changing in all of its forms. Besides, there's not all that many places where such a need would arise - if an awkward cluster forms, one of the gzarot will take over and an extra vowel inserted.


There's only one place in Hebrew where metathesis occurs that I know of, but it's not with the root, but with a root letter and a prefix. It occurs in the past tense of binyan hitpa'el verbs:

In the past tense of hitpa'el, the prefix hit- is added to the root in all of its forms (along with vowel changes and suffixes at the end relaying person/number/gender information). The regular hitpa'el past tense pattern looks like this (with the example root *Y-SH-B "settle" in parentheses):

1Sg: hitCaCaCti (hityashavti)
2SgM: hitCaCaCta (hityashavta)
2SgF: hitCaCaCt (hityashavt)
3SgM: hitCaCeC (hityashev)
3SgF: hitCaCCah (hityashvah)
1Pl: hitCaCaCnu (hityashavnu)
2PlM: hitCaCaCtem (hityashavtem)
2PlF: hitCaCaCten (hityashavten)
3Pl: hitCaCCu (hityashvu)

However, if the first letter of the root is one of the so-called 'whistling consonants' ('otiyot shorqot - that is, /s z S tz/), the t of the hit- prefix and the first root letter metathesize (and sometimes a few other languages occur, but that's not important here). This pattern, Gizrat P"[otiyot shorqot], is shown here in the past tense with the example root *S-K-L 'to look at'):

1Sg: hiCtaCaCti (histakalti)
2SgM: hiCtaCaCta (histakalta)
2SgF: hiCtaCaCt (histakalt)
3SgM: hiCtaCeC (histakel)
3SgF: hiCtaCCah (histaklah)
1Pl: hiCtaCaCnu (histakalnu)
2PlM: hiCtaCaCtem (histakaltem)
2PlF: hiCtaCaCten (histakalten)
3Pl: hiCtaCCu (histaklu)
Last edited by Mecislau on Wed Aug 22, 2007 7:31 pm, edited 1 time in total.

User avatar
Matt
Lebom
Lebom
Posts: 138
Joined: Mon Mar 15, 2004 2:11 pm
Location: Kansas

Post by Matt »

I have a question about triconsonantal root systems that's been bugging me for some time. Let's say that there are two roots, K-R-S and R-S-T. Now let's posit the word karasata, and assume that ka- and -ta are valid affixes.

Are there any comparable ambiguities in real-life triconsonantal root languages? I suppose that context would make it clear, but what about in isolation?
Kuku-kuku kaki kakak kakekku kaku kaku.
'the toenails of my grandfather's elder brother are stiff'

User avatar
Mecislau
Avisaru
Avisaru
Posts: 491
Joined: Thu Jul 24, 2003 2:40 pm
Location: Maryland
Contact:

Post by Mecislau »

Matt wrote:I have a question about triconsonantal root systems that's been bugging me for some time. Let's say that there are two roots, K-R-S and R-S-T. Now let's posit the word karasata, and assume that ka- and -ta are valid affixes.
Of course that could happen. It's just like any language - some words can have what looks like a grammatical affix, but isn't. Look at all of the singular English nouns ending in /s/, for example. We know when it's a plural noun or a singular one that just looks plural (with some exceptions - the English word "pea" came from an older word "pease", which was both singular and plural, but some people confused the final /s/ for a plural suffix, giving us a singular "pea").
Matt wrote:Are there any comparable ambiguities in real-life triconsonantal root languages? I suppose that context would make it clear, but what about in isolation?
None come to mind right off hand, but I'm sure there's plenty.

User avatar
blank stare II
Lebom
Lebom
Posts: 127
Joined: Sat Jun 18, 2005 4:34 pm
Location: second to the right and straight on till morning
Contact:

Post by blank stare II »

How often do such systems arise? And do they often consist of different numbers of radicals(like two or four)? Are the Semitic languages the only ones which do this?
I get a big kick out of playing my own language game–it’s a unique thrill only conlangers know.
- J Burke

Grath
Sanci
Sanci
Posts: 20
Joined: Sun Oct 24, 2004 8:09 am
Location: Site of the Great PIE *a Theft

Post by Grath »

I think other Afro-Asiatic languages have biconsonental root systems, don't they?

User avatar
Mecislau
Avisaru
Avisaru
Posts: 491
Joined: Thu Jul 24, 2003 2:40 pm
Location: Maryland
Contact:

Post by Mecislau »

Grath wrote:I think other Afro-Asiatic languages have biconsonental root systems, don't they?
Proto-Afro-Asiatic did, though I'm not sure about the other AA languages. It'd make sense, though, if some of them are.


I've also heard that PIE was sort of heading toward developing a biconsonantal system, but it never came about. (Remember, ablaut, such as English or German's, is not the same as a biconsonantal system)

User avatar
Mecislau
Avisaru
Avisaru
Posts: 491
Joined: Thu Jul 24, 2003 2:40 pm
Location: Maryland
Contact:

Post by Mecislau »

Y'know what, I feel like reviving this thread.


First of all, one little interesting note: Modern Hebrew is currently in the process of rapidly expanding its quadraconsonantal roots. It's forming new ones. How?

Through a process much like one English underwent, actually. Consider the following cycle, where a verb gave rise to a noun, that in turn gave rise to a new verb:

profess > profession > professional > professionalize

As well as many other verbs, like "condensate" (from condensation, instead of condense) or "obligate" (from obligation, instead of oblige).


Well, remember how I said earlier how Hebrew extracted the consonants from many foreign loan words to make new verbs, like how the loan טלפון telefon gave the root T-L-P-N, which in turn gave the new verb מטלפנת metalpenet "I/You/She is telephoning..." and all of its forms?

One of the nominalization patterns Hebrew uses is the "tool" pattern, whose shalem pattern is maCCeC. Ex: מחשב machshev "computer" from the root ח-ש-ב (CH-SH-B 'to think'). Hey, wait a minute! מחשב has four consonants, despite being derived from a triconsonantal root through a regular pattern. Well, if it has four consonants, what's to keep me from doing with it the same thing that happened to telefon?

And so the new root M-CH-SH-B was born, meaning "to computerize".

היא ממחשבת את הרשומות עיסקיות שלחם
hi' memachshevet 'et harshumot ha'isqiyot shelachem
"She's computerizing your business records"



Also, I recently got a book that includes several chapters about how exactly the triconsonantal system came into existance from what was originally a very Indo-Europeanesque Proto-Afro-Asiatic grammar. Is anyone interested in a summary?

Rory
Lebom
Lebom
Posts: 226
Joined: Sun Jun 15, 2003 4:37 pm
Location: Scotland
Contact:

Post by Rory »

I wonder if I can work some of this stuff into a IE-style language family...
The man of science is perceiving and endowed with vision whereas he who is ignorant and neglectful of this development is blind. The investigating mind is attentive, alive; the mind callous and indifferent is deaf and dead. - 'Abdu'l-Bahá

User avatar
alice
Avisaru
Avisaru
Posts: 707
Joined: Wed Oct 30, 2002 4:43 pm
Location: Three of them

Post by alice »

Rory wrote:I wonder if I can work some of this stuff into a IE-style language family...
That's exactly what I was thinking of.
Zompist's Markov generator wrote:it was labelled" orange marmalade," but that is unutterably hideous.

User avatar
Trebor
Lebom
Lebom
Posts: 207
Joined: Sat Apr 10, 2004 7:36 pm
Location: Canada

Post by Trebor »

Maknas wrote:Also, I recently got a book that includes several chapters about how exactly the triconsonantal system came into existance from what was originally a very Indo-Europeanesque Proto-Afro-Asiatic grammar. Is anyone interested in a summary?
/me raises his hand wildly, shouting, "Me! Me! Me! As soon as fucking possible!" (You shall see sooner or later why.)

User avatar
Mecislau
Avisaru
Avisaru
Posts: 491
Joined: Thu Jul 24, 2003 2:40 pm
Location: Maryland
Contact:

Post by Mecislau »

Rory wrote:I wonder if I can work some of this stuff into a IE-style language family...
Of course you could. As I said, reconstructed Proto-Afro-Asiatic looks very IE-esque morphologically. And phonetically. And in many other respects... :?


Anyway, I'll try to write up a description as soon as possible (if it's not ready by this weekend due to my laziness, yell at me. Loudly). It basically involves a whole lot of syncope, affixation, reanalysis, analogy, and laryngeals screwing up vowels. Especially the analogy and syncope parts.
Last edited by Mecislau on Wed Aug 22, 2007 7:33 pm, edited 1 time in total.

Ran
Lebom
Lebom
Posts: 145
Joined: Fri Sep 13, 2002 9:37 pm
Location: Winterfell / Lannisport / Highgarden
Contact:

Post by Ran »

I have a question... what happens when sound changes occur?

Like for example, when consonants palatalize before a front vowel? (ki to chi)

Nasals lengthen the previous vowel, then disappears? (ant to a:t)

Plosive clusters assimilate to each other, to become long plosives? (pt to tt)

Consonants drop out at the end of a word?

Lenition in intervocalic environments? (t becomes d, but only between two vowels)

At the rate that these happen in the two groups of languages that I'm slightly familiar with (western IE langs and Sinitic), a triconsonantal root language would get pretty messed up within about 1000 to 2000 years. Or less.

Or do these changes not happen at all in AA? (The paradigms that you've shown seem pretty regular...) What kind of sound changes do happen then, and can you tell me the effects that they have on this entire system?
Winter is coming

User avatar
Debegduk ing Debegduked
Sanci
Sanci
Posts: 16
Joined: Fri Dec 03, 2004 11:12 pm

Post by Debegduk ing Debegduked »

Matt wrote:I have a question about triconsonantal root systems that's been bugging me for some time. Let's say that there are two roots, K-R-S and R-S-T. Now let's posit the word karasata, and assume that ka- and -ta are valid affixes.
Reminds me of what happened when Swahili borrowed the word kitab "book" from Arabic. There's a Swahili noun class that uses ki-/vi- for plural and singular, and kitab got re-analysed to fit into that class, making the Swahili plural vitab, to the bemusement of the Arabs.
U bronazda Lin, broZuspabimaded guze ner liuringemeni!

User avatar
johanpeturdam
Sanci
Sanci
Posts: 36
Joined: Sat Nov 06, 2004 9:32 pm
Location: Bratislava, Slovakia, originally: Funningur, Faroe Islands
Contact:

Post by johanpeturdam »

geoff wrote:
Rory wrote:I wonder if I can work some of this stuff into a IE-style language family...
That's exactly what I was thinking of.
Stop me, if I'm wrong, but I'm pretty sure, that I've read or heard, that Persian has a tri-consonant system. If it's fully developed I'm far from able to answer.

You have mentioned bi-consonant systems in PIE. But are you sure, that there aren't traces still in IE languages?

Faroese example:

barn (child)
bera (to carry)
byr?a (burden)
bur?ur (carry load)

OK, these are only 4 words, but b-r is in all of them. And all involve carrying. In Faroese a way to say a woman is pregnant, is to say that "hon ber barn undir belti" (lit. she carries child under belt). Does this work as sort of a hint of a possible bi-consonant system in PIE?
Ungur nemur, gamal fremur
Da giovani si impara, da adulti si applica

User avatar
Mecislau
Avisaru
Avisaru
Posts: 491
Joined: Thu Jul 24, 2003 2:40 pm
Location: Maryland
Contact:

Post by Mecislau »

Ran wrote:I have a question... what happens when sound changes occur?
Well, first of all, yes, the roots have some very powerful analogy 'guarding' them from the havoc that would be created by sound changes. Changes do occur, but they're much slower, and have a greater tendency to be all-encompassing (eg, no restrictions). Even if T turns to D intervocally, it's always possible that other Ts could become D analogically.

Some of the changes you've mentioned already exist in Hebrew. There's a series of lenited consonants, where k > /X/, p > /f/, and b > /v/ according to a mind-numbingly series of rules. Now, Hebrew already had /X/ and /v/ phonemes, so there's a new merger here (sort of... technically the /X/ and the /X/-from-/k/ behave slightly differently within the system, but... yeah, nevermind). Now, you could just consider X, f, and v to be allophones of their original consonants. Or, you could consider them allomorphs - phonemes in alteration with other phonemes according to certain morphological rules.

Also, Hebrew's gzarot (irregular paradigms) exist because of past sound changes that threw some roots off of the regular pattern. For example, in the binyan hif'il, all initial /n/ in roots has been completely elided. So, for example, "I know" (as in, familiar with), which has the root N-K-R, is "makir", instead of *mankir. This has just succeeded in creating a new irregular paradigm, known as gizrat P"N, where the N of the root drops in the present tense and infinitive. Predictably.


johanpeturdam wrote:barn (child)
bera (to carry)
byr?a (burden)
bur?ur (carry load)
That is IE ablaut, a system related, but much less extensive or powerful as the triconsonantal roots.

A single Hebrew root can yield dozens of seperate words, all according to regular patterns.

Plus, the patterns can apply to any root. You say all of those Faroese words contain B-R. Well, can you plug in any two consonants, say, some meaning "read", and end up with CaCn "reader", CeCa "to read", CyC?a "book", etc?
Last edited by Mecislau on Wed Aug 22, 2007 7:37 pm, edited 1 time in total.

User avatar
alice
Avisaru
Avisaru
Posts: 707
Joined: Wed Oct 30, 2002 4:43 pm
Location: Three of them

Post by alice »

Debegduk ing Debegduked wrote:
Matt wrote:I have a question about triconsonantal root systems that's been bugging me for some time. Let's say that there are two roots, K-R-S and R-S-T. Now let's posit the word karasata, and assume that ka- and -ta are valid affixes.
Reminds me of what happened when Swahili borrowed the word kitab "book" from Arabic. There's a Swahili noun class that uses ki-/vi- for plural and singular, and kitab got re-analysed to fit into that class, making the Swahili plural vitab, to the bemusement of the Arabs.
Actually those are kitabu and vitabu in Swahili, in which all syllables are open.
Zompist's Markov generator wrote:it was labelled" orange marmalade," but that is unutterably hideous.

User avatar
Drydic
Smeric
Smeric
Posts: 1652
Joined: Tue Oct 08, 2002 12:23 pm
Location: I am a prisoner in my own mind.
Contact:

Post by Drydic »

Ran wrote:I have a question... what happens when sound changes occur?

Like for example, when consonants palatalize before a front vowel? (ki to chi)

Nasals lengthen the previous vowel, then disappears? (ant to a:t)

Plosive clusters assimilate to each other, to become long plosives? (pt to tt)

Consonants drop out at the end of a word?

Lenition in intervocalic environments? (t becomes d, but only between two vowels)

At the rate that these happen in the two groups of languages that I'm slightly familiar with (western IE langs and Sinitic), a triconsonantal root language would get pretty messed up within about 1000 to 2000 years. Or less.

Or do these changes not happen at all in AA? (The paradigms that you've shown seem pretty regular...) What kind of sound changes do happen then, and can you tell me the effects that they have on this entire system?
I got a copy of Chris Ehret's Reconstructing Proto-Afroasiatic a few weeks ago, and it seems that, for the proto-langs at least, don't have those sort of sound changes. The other AA branches besides Semitic and Egyptian (to an extent, and it has nothing useful about Berber) don't have the system quite so prevalent. It's still there, but nowhere near as strongly.

As for the PAA being PIElike, yes, except it seems to have prefixing grammatical units far more frequently than suffixal. I may post some examples later.
Image Image
Common Zein Scratchpad & other Stuffs! OMG AN ACTUAL CONPOST WTFBBQ

Formerly known as Drydic.

User avatar
Noriega
Sanci
Sanci
Posts: 48
Joined: Tue May 04, 2004 7:41 am
Location: Skaane

Post by Noriega »

Maknas wrote:One of the nominalization patterns Hebrew uses is the "tool" pattern, whose shalem pattern is maCCeC.
The Arabic tool/instrumental pattern is miCCaaC. The only word I can remember right now is miftaah "key". It is based on the pattern F-T-H, meaning "to open". However, the Icelandic word for key is lykill, meaning "something that locks". Well, that was just your random language diversity comment ;)
Perhaps eventually all languages will evolve so that they include some clicks among their consonants – Peter Ladefoged

Jahai: /kpotkpɛt/ ‘the feeling of waking up to the sound of munching’

User avatar
Mecislau
Avisaru
Avisaru
Posts: 491
Joined: Thu Jul 24, 2003 2:40 pm
Location: Maryland
Contact:

Post by Mecislau »

Noriega wrote:
Maknas wrote:One of the nominalization patterns Hebrew uses is the "tool" pattern, whose shalem pattern is maCCeC.
The Arabic tool/instrumental pattern is miCCaaC. The only word I can remember right now is miftaah "key". It is based on the pattern F-T-H, meaning "to open". However, the Icelandic word for key is lykill, meaning "something that locks". Well, that was just your random language diversity comment ;)
Heh, yeah. Hebrew does a similar thing: מפתח maft?ach, from the root P-T-CH "open".



Also, Drydic, that book is excellent. I've gotten it from the library before, and it's quite an interesting read. It's also where I got most of the Afro-Asiatic sound changes that I added to the Correspondence Library. If only they'd use a more consistent transcription...

User avatar
johanpeturdam
Sanci
Sanci
Posts: 36
Joined: Sat Nov 06, 2004 9:32 pm
Location: Bratislava, Slovakia, originally: Funningur, Faroe Islands
Contact:

Post by johanpeturdam »

Maknas wrote:Plus, the patterns can apply to any root. You say all of those Faroese words contain B-R. Well, can you plug in any two consonants, say, some meaning "read", and end up with CaCn "reader", CeCa "to read", CyC?a "book", etc?
No, but the point of my post, was just to know, if those examples would qualify as an indicator in a pre-Germanic or even proto-Indo-European bi-consonant root system. Faroese doesn't use bi/tri-consonant root systems.

But since you asked:

L-S

lesa (to read)
lesari (reader)
lestur (reading (noun))

But seriously, aren't there still traces in modern I-E languages, for instance with that B-R system I showed, which show a possible bi/tri-consonant root system? And if this IS the case, then is this also why some linguists want to tie the I-E languages and A-A languages together under one family?
Ungur nemur, gamal fremur
Da giovani si impara, da adulti si applica

User avatar
Noriega
Sanci
Sanci
Posts: 48
Joined: Tue May 04, 2004 7:41 am
Location: Skaane

Post by Noriega »

johanpeturdam wrote:L-S

lesa (to read)
lesari (reader)
lestur (reading (noun))
Isn't this just "lesa" with suffixes, and a disappearning -a in the last word?

But I also believe that there is some kind of consonant stems in Scandinavian, and your B-R examples are better. Here's a little page I once made on the subject:

http://home.unilang.org/main/wiki2/inde ... nant_stems
Perhaps eventually all languages will evolve so that they include some clicks among their consonants – Peter Ladefoged

Jahai: /kpotkpɛt/ ‘the feeling of waking up to the sound of munching’

User avatar
Mecislau
Avisaru
Avisaru
Posts: 491
Joined: Thu Jul 24, 2003 2:40 pm
Location: Maryland
Contact:

Post by Mecislau »

Alright, part one of the origin of Semitic triconsonantal roots...



The ancestor of the Semitic languages (pre-Proto-Semitic, we could say) had perfectly pronounceable roots, much like IE languages. The three my book uses as examples are *ktum 'cover', *ptil 'twist', and *mūt 'die'.

Let's use work in the first person. The first person singular past was formed by adding a- to the root:

aktum "I covered"
aptil "I twisted"
amūt "I died""

(Why are we working in the past? Well, Proto-Semitic didn't have a present tense - it worked within an imperfective/perfective system. The Hebrew present tense derives from earlier participles, which is why the present tense verbs act more like adjectives than the other verbs).

The future tense was formed by adding some ending to the above past forms. This ending is long gone, so let's just call it -ā. So the future tense would be:

aktumā "I will cover"
aptilā "I will twist"
amūtā "I will die"

From here, PSS took the route of the Germanic languages. The ending -ā caused the vowel immediately preceding it (ie, the vowel of the root) to lower to /a/, while preserving the original length:

aktamā "I will cover"
aptalā "I will twist"
amātā "I will die"

The future tense ending then dropped. We now have ablaut distinguishing the past tense from the future tense:

aktum > aktam
aptil > aptal
mūt > amāt

(Now, it's possible that the assimilation didn't occur for all roots, but analogy would have spread the ablaut alteration)

Now we've got the very beginnings of the consonantal root beginning to appear. Vowel alterations are now important, but this is still nothing special - the Germanic languages do more than this.

We've got a few verbs here with three consonants, but like English, they're still connected to each other. Just like you can say "spin", "span", and "spun" by changing the vowel, the "sp" are still stuck together. The same is true with the "kt" or "pt" in the roots above: the "kt" and "m" in "aktVm" are seperate, but the k and t are stuck. We need to seperate these consonants first.

Now for a short diversion that actually is relevant, despite how it may seem at first:

PSS at this point had many roots that contained both three consonants (like ktum and ptil) and only two consonants (like mūt). Modern Semitic languages are majority triconsonantal. So how did all of these biconsonantal roots become triconsonantal?

Affixes. Affixes that fused to the root. For example, let's have a verb *pil, meaning "lie" (as in lie down). To turn this into an adjective, the PSS speakers added some sort of prefix. Here they could add *?a-, meaning "low", to form the adjective *?apil "low-lying". PSS also had a case system for adjectives, so let's add the ending -um (it doesn't matter what it means in this example), to get the affixed *?apilum.

Remember how I mentioned earlier the cycle of verb to [some other part of speech] to verb again, such as "compute" > "computer" > "computerize"? Well, the same process could turn the adjective *?apil back into a verb: *a?apil "I became low" (and, of course, its future form *a?apal by analogy).


Now, back the formation of purely consonantal roots.

We've got a simple ablaut system going, but what happens when you introduce a powerful syncope? One occured in PSS, which eliminated every penult vowel in words with three or more syllables. So, how does this affect our verbs?

*aktum, *aptil, and *amūt are unaffected, being only two syllables. But what about our adjective *?apilum and our verb *a?apil? They would be shortened:

?apilum > ?aplum
a?apil > a?pil

The verb now looks a lot more like the older verbs.

But wait a minute! Let's look at the structure of these two words. The adjective is CaCCum. The verb is aCCiC. The consonants have all been split apart! In the adjective, we've got ? and pl paired, while in the verb we have ?p and l.

Analogy then kicks in:

The verb *a?pil "I became low" yielded the adjective *?aplum "low" by rearranging the consonants and vowel from aCCiC to CaCCum. The verb *aptil "I twisted" has the same structure as *a?pil. So, what would its adjective form be? *patlum "twisted", of course!


And so all three consonants were liberated from each other. In the speakers' minds, the vowels were no longer an important part of the stems, and so we now have purely consonantal roots. From here, we just need to develop new vowel patterns...




That'll be for next time.





(Information/Theories are from "The Unfolding of Language" by Guy Deutscher).
Last edited by Mecislau on Wed Aug 22, 2007 7:40 pm, edited 1 time in total.

User avatar
Mecislau
Avisaru
Avisaru
Posts: 491
Joined: Thu Jul 24, 2003 2:40 pm
Location: Maryland
Contact:

Post by Mecislau »

First of all, I called the cause of the original tense ablaut the prefix -ā. Of course, that's only one theory. Another common one is the Laryngeal Theory (fun, Semitic gets one too)

In the modern Semitic languages, it is clear laryngeals like to mess up vowels a bit. They tend to drag vowels closer to /a/, because /a/ is the closest vowel shape-of-tongue-wise to most laryngeal consonants. For example, look at the Hebrew noun תפוח, originally tapuch, but the final /X/ caused the vowel to break, yielding modern tapúach. Note that this only happens at the end of a word: the plural is still regular, תפוחים tapuchim.

So, let's say instead that the past tense was formed by adding an ending (any one, it doesn't matter) to the present/future form. Let's just stick with -ā. So we've got (when adding in a new laryngeal-final root, *şīħ 'laugh')

Root - Past - Future
ptil - aptilā - aptil
mūt - amūtā - amūt
şīħ - aşīħā - aşīħ

Now, let's have a change like in Hebrew: the word-final laryngeal causes the preceding vowel to break and introduce an /a/. So, the last one becomes: şīħ - aşīħā - aşīaħ.

Now, we've got speakers who begin to analagize this introduced /a/ to all verbs, thinking that it's an integral part of the future tense. This vowel then overtakes the former vowel, leaving:

Root - Past - Future
ptil - aptilā - aptal
mūt - amūtā - amāt
şīħ - aşīħā - aşāħ

And then the final -ā marking the past tense drops through regular sound change, giving us our new ablaut system once again:

Root - Past - Future
ptil - aptil - aptal
mūt - amūt - amāt
şīħ - aşīħ - aşāħ



Anyways...

Alright, so we've got tenses and an adjective. But what about that mess of binyanim, where a verb can take a slightly diffeent vowel pattern to acquire a new meaning? How did those form? Let's look at a few specific patterns...

The Reflexive: Doing Something to Oneself

The Hebrew Reflexive Pattern: hitCaCeC (Third Person Singular Masculine Past Tense Reflexive)
The Arabic Reflexive Pattern: iCtaCiC (Imperative Reflexive)

The reflexive in most Semitic languages is made by inserting a "t" in between the first and second consonants, or, in the case of Hebrew, before the first. This likely came from an ancient reflexive pronoun, something like ta, which eventually prefixed itself to the verb. In most Semitic languages, it metathesized with the first consonant of the root, yielding their modern patterns.

In Hebrew, the metathesis wasn't as overarching as in Arabic. The hit- prefix remains in front of the root. However, if the first consonant of the root is /s/, /z/, /ts/, or /S/, the metathesis does occur: *S-K-L "look at, watch" becomes הסתכל histakel, instead of the usual *hitsakel.

The Intensive - Doing Something Intensely

Hebrew Intensive Pattern: CiCCeC (Third Person Masculine Singular Past Tense Intensive)
Akkadian Intensive Pattern: uCaCCaC (First Person Singular Future Tense Intensive).

As you can see, the intensive is usually formed by geminating the second consonant. In Modern Hebrew, however, the geminate has collapsed, making the vowels more important.

This was likely originally formed by reduplication, as many languages reduplicate a word to make it more intense.

(Conversely, modern Hebrew does the same to make an adjective less intense: ירוק yaroq "green", ירוקרוק yeroqroq "greenish". This is a form of diminutive.)

You don't have to reduplicate the whole word: just one syllable is fine. So, if I may use the root *D-B-R "talk" (which is now limited to the intensive pattern due to blurring of meaning)...

An older *daber could have been made intensive by reduplicating to *debdaber. Assimilation then changed this to *debbeber. Then syncope and assimilation again reduces this to *dibber.

The Causative - To Make Someone Do Something

Hebrew Causative Pattern: hiCCiC (Third Person Masculine Singular Past Tense Causative)
Akkadian Causative Pattern: ushaCCiC (ditto; the u- is a pronomial)

This is formed by adding sha-, or a reduced form ha-, to the root. There are two possible origins:

Like the Reflexive, this could have been its own word at one point, likely meaning something like "make" or "cause". The same thing happened in English, although it didn't expand that far: in Proto-Germanic, the verb "ian" (make) attached to the other of other verbs to make causatives: "fallian" (to cause to fall), "drankian" (to cause to drink) (Don't shoot me - I don't know Proto-Germanic, so I made up some of the forms... this change did actually happen, though).

This suffix caused the vowels to change, and then the suffix dropped. As a result, English is left with a few causative pairs, such as:

to fall ~ to fell
to drink ~ to drench
to sit ~ to set
to rise ~ to raise

Etc.



However, this likely didn't happen in Semitic. Remember, I mentioned the sha- prefix earlier - it turned verbs into adjectives. So, we could have a cycle like:

*pil "lie" > *sha-pil "low-lying" > *shapil "to make low-lying" (ie, "to cause to lie down").

Hey, look! If we ignore the adjective in the middle, since speakers aren't going to remember the order of events generations after they happen, we have a causative: "lie" became "to cause to lie". So sha- could be reananlyzed as a causative.


The Passive - To Be Done

The passive originated from a prefixed n-, likely deriving from a verb meaning "to be" or "to become" followed by an adjective. In Hebrew, the basic passive pattern is niCCaC.

However, there is more than passive possible. I'm not exactly sure how these others formed... Posit your own hypotheses! :P

For example, Hebrew has two passive binyanim: A passive intensive and a passive causative:

Active | Passive

Intensive: CiCCeC, CuCCaC
Causative: hiCCiC, huCCaC

As you can see, the passive was formed by taking the active pattern and changing the vowels to u-a. How n- caused this, I'm not sure... Mabye the prefix was more complex at first?


Binyan Formation in Action - Or How Politics Affects Language

Well, that's fine 'n dandy. We've got several hypothetical examples. But can we actually trace any of them?

Yes, actually. Modern Hebrew colloquially has developed an eighth binyan, called hitpu'al, which is the passive of reflexive: "to be made to do something to oneself". It was created in the 1940s :P

This sounds a bit ridiculous at first, but it does have a use. For example, the root *P-T-R "set free" in the regular reflexive means "set oneself free", or by extension, "to resign". Now, it's perfectly likely to talk about someone being made to resign themselves...

Plus how about "to be made to wash oneself"? Or "to be made to volunteer oneself"? Those are relatively useful.

Anyway, it's origins. In 1948, an Israeli politician was forced to resign (you know how it goes...). Being very frustrated, he coined a new binyan in a speech expressing what happened, by analogy. Let's look at this:

Active | Passive
Intensive: piter "he set free (ie, he fired somebody)", putar "he was set free (ie, he was fired)"

Causative: hiftir "he caused somebody to be set free (ie, he caused somebody to be fired), huftar "he was cause... you get the point)

So, piter > putar, hiftir > huftar...

Now we have hitpater "he set himself free/resigned". Well, both of those passives are formed by using the vowel pattern u-a, so...

Hitputar! "He was made to set himself free"/"He was resigned"

Of course, this is nonstandard, but shows the powerful force of analogy.





And if you were curious, here's the root *P-T-R in each of the binyanim (all 3 Sg Masc Past), for meaning comparison:

Pa'al (Default): פטר patar "he dismissed, he exempted"
Nif'al (Simple Passive): נפטר niftar "he was dismissed, he was released, he passed away"

Pi'el (Intensive): פיטר piter "he fired, he discharged"
Pu'al (Passive of Intensive): פוטר putar "he was fired, he was discharged"

Hif'il (Causative): הפטיר hiftir "he caused smby to be fired"
Huf'al (Passive of Causative): הופטר huftar "he was caused to be fired by somebody"

Hitpa'el (Reflexive): התפטר hitpater "he resigned, he got rid of somebody"
*Hitpu'al (Passive of Reflexive): התפוטר hitputar "he was made to resign, he was gotten rid of"



Okay, that's it for now. From here it should be easy to see how this system can be expanded to include all sorts of concepts.
Last edited by Mecislau on Wed Aug 22, 2007 7:44 pm, edited 1 time in total.

Post Reply