11 posts tagged “sanskrit”
A while ago I posited the idea that the paradigm for 'foot' in Indo-European might have a phonotactic schwa insertion. Glen Gordon spook out against this, saying it was paradigmatic levelling rather than a phonotactic constraint. I didn't want to believe him, simply because phonotactic constraits are so much nicer than paradigmatic levelling, but I found some examples which shows that he's probably right.
So let us look at the paradigm for 'foot'. We have the root *ped- With a paradigm like this:
Nom. *pṓd-s
Gen. *ped-ós
The ō in the nominative is due to mono-syllabicity of the word. The vocalism of this word is then perfectly acceptable.
The genitive though is slightly more problematic. From the idea of syncope, we'd rather see a word like **pd-ós, which we clearly don't. My idea was then that a *e was inserted (which was previously a schwa), due to phonotactic constraints, after all /pd/ is a rather difficult cluster to pronounce initially, even Greek, known for it's sometimes rather strange initial clusters doesn't allow such a cluster. Glen Gordon argued that this is not so much a phonotactic constraint, but rather the tendency of Indo-European to avoid asyllabic roots alternating with syllabic roots in one paradigm.
At first I did not see why one would assume this, and then I suddenly remembered the root *peth₁- 'to fly'. This root is well attested in many languages, including English (feather). But the Greek word πτερόν 'wing' is the one I'd like to talk about. As we can see, Greek has a perfect zero grade in the root for this word, a Indo-European reconstruction of this word would be:
*pth₁ró-m
By the phonotactic constraint I first proposed such a cluster would be forbidden. I would expect a form **peth₁ró-m, so I have to re-evaluate my theory. I'm not quite feeling into giving into Glen's theory yet, and would like to change my theory a bit;
When there's a initial cluster of two consonants of which one of the two elements is voiced, a schwa is insterted between the two.
I believe Greek agrees with me on this one, since there's no clusters like bd, gd, pd, bt etc. to be found in the Greek language initially. Nevertheless I'm going to have to search for words to disprove me again; If you're feeling in a particularly productive mood, I encourage my readers to do the work for me!
Just a little side not to finish this post of, it might be interesting to know that Sanskrit in fact does have the phonotactic constraint that I posited the first time, the word for 'wing' in Sanskrit is patram पत्रम्. Which is a problematic word as well since it brings up some of Sanskrit's sporadic laryngeal vocalisation, but we'll leave that for some other time.
Sanskrit does allow such clusters medially though. The third person plural perfect in Vedic Sanskrit of patati पतति 'to fall/fly' is paptuḥ पप्तुः While in Classical Sanskrit by a bizarre analogy from sad- 'to sit' it became petuḥ पेतुः. To sit in 3ppl perf. is seduḥ < *sazduḥ < *sasduḥ. This e-vocalism rather than a medial CC cluster was generalised to all reduplicating perfects that have identical initial reduplication (Labials, dentals and fricatives). Seems like Sanskrit wasn't particularly fond of the medial zero-grade roots after a while either.
What I've always found fascinating about Indo-European is that there seems to be many isoglosses that alternate certain vowels. *e/o alternations in, for example the genitive *-es/-os. Then there's the *i/e alternation in reduplications. Where we find i in Present/Aorist reduplications, and e in perfect reduplication (although we find many variants of the reduplication where the root vowel seems to influence the reduplicated vowel). And last but not least there's the *i/u alternation, which is what I'll talk about today.
I have no idea what may be the solution to this odd alternation, but I thought I'd bring it up to our collective awareness.
Sanskrit has the locative plural -su सु/षु. This is apparently related to the Greek dative plural ending -σι(ν). As you can see u/i seem to alternate. But this doesn't always happen between Sanskrit and Greek. For example the prefix *dus- seen in both Greek and Sanskrit both display proof for an original *u. Sanskrit has the prefix dus- दुस्- And Greek has δυσ-. For example in Sanskrit you have the word durjanaḥ दुर्जनः 'a bad man' which is quite similar to the greek word: δυσγένεια 'low of birth'.
But then if we look at the Latin equivalent of this prefix we find dis-! It's the dis- we find in dislocation for example. So somehow the *u/i alternation does not seem to be a steady, or split by isoglosses that can be strictly defined in the modern day languages.
I applaud anyone willing to come up with an explanation to this bizarre alternation. I for one, do not have a clue where it came from.
For some time I've been wondering about the Indo-European genitive forms; Which anyone who reads my blog with any regularity has surely noticed.
As some of you may know, Indo-European has a different genitive form the nominal inflection than for the pronominal inflection. Later on the o-stems have taken up a lot of features of the pronominal inflection, as well as the pronominal genitive, but Hittite has convincing proof that this was not always the case; there Nominative and Genitive of a-stems (< PIE o-stems) are identical, ending in -aš.
The Indo-European pronominal ending is *-eso or *-oso; while the nominal ending is *-os or *-es. I believe I have found a way to unify these two forms as an originally identical ending.
To explain this I'll first have to bring in some theories which Glen Gordon has been working on considering a Pre-Indo-European syncope. Let me just stress that everything I present here are my theories, I'm not doing this to get respect and honour, but just so people won't think badly of Glen Gordon's awesome work if I make some terrible mistake in my reasoning.
Now for some spelling conventions. I am fairly traditional in displaying *e as *e and *o as *o, while it is quite likely that in fact the *o was an *a at some earlier stage. For an unknown vowel in my reconstructed pre-Indo-European forms I will use *a. Well, on to the actual theory now then!
For a word like 'father' we find the following paradigm
Nom. *ph₂tḗr
Gen. *ph₂trós
Now considering the 'Quasi-Penultimate Accent Rule' (QAR) and the syncope of almost all, if not all vowels outside of the accent we'd expect a Pre-Indo-European paradigm like this:
Nom. *p(a)h₂téro=so With pseudo-suffix *so-/to- 'this'
Gen. *p(a)h₂teró-sa With a true suffix and thus agreeing to the QAR.
The *so suffix later gave rise to a *rs cluster which lost the *s with compensatory lengthening, hence *ḗ. A similar process is later seen in classical Greek.
The syncopated final letter *o which appears in the genitive could also be the vowel *e, which gives a very nice explanation why most languages have generalised the *-ós suffix, while some (like Latin) show an *-és suffix.
Having shown this, let's now get to the really interesting part, the pronouns.
The previously mentioned stem *so-/to- 'this'(with it's curious alternation of stem *so- in the nominative *to- in the oblique) is great to show as an example.
Nom. *so(s)
Gen. *toso
The *-s ending in the nominative is obviously highly curious if it is indeed from the stem *so-/to- because you'd be saying *so=so 'this-this' in Pre-Indo-European. Luckily there's some proof that this is not the original form. Greek has the form ὁ which goes back to *so, and definitely not to *sos. Also there's the Sanskrit form of this word सः saḥ which has a variant 'Sandhi' form स sa. Although this variant isn't usable in all contexts, the form is not a regular sandhi variant, which makes it likely that this is in fact the older form.
I think with that I have sufficient proof to not reconstruct this nominative *-s for pre-Indo-European
Now onto the pre-Indo-European forms
Nom. *so
Gen. *to-so
What can be said about these forms?
First, the nominative taking the penultimate accentuation into account, should be accented before itself, which is not even unthinkable (accentless words in Greek can do this). How come it didn't syncopate then? Well, that can have several reasons. Either the form *so was in fact *só dealing with the impossible accentuation by simply placing the accent on the only possible position. We do not have much proof for this though. It is in fact likelier that this particle was just 'unaccented'. How come it didn't undergo syncope? Well the fact that the word would've then be **s might be an indication. Another reason is that really short particle-like words like *so-/to- tend to be very resistant to such syncopations which normal words regularly take part in.
Although the Genitive could in fact have been accentuated as **tóso giving a later form like **tós it clearly didn't. An explanation for this might be that the stem *so-/to- wasn't accentuated in the nominative and was thus felt as an 'unaccentable word'. Besides that this is of course still one of these particle-like words, even though it's in the genitive. Therefore not taking an accent but still being strangely resistant to syncopation.
Now, if we put next to each other the reconstructed genitive of father, and that of this we see the following:
Gen. *p(a)h₂teró-sa
Gen. *to-so
Where in the first form the suffix' vowel was still a mystery vowel, from the pronominal form it is quite clear that this should be *-o thus giving:
Gen. *p(a)h₂teró-so
With this the two forms of the genitive have finally been unified into a single suffix *-so. So obvious that I'm astounded it hadn't occurred to me sooner. Besides that it's also so obvious I'm sure it has occurred to other people too. nevertheless I'm here to inform both myself and my readers, so even if I'm the last one to discover this, it's still useful for me.
I promised a while ago that I'd be writing down the phonology of Hittite on this blog. I am still at it, it's turning out to be enormous. But please be patient, while you wait I'll have some musings on the origin of the Indo-European wolf.
Wolf is a very odd word, because finding the 'right' root is actually far more difficult than you'd think. Traditionally we reconstruct *ul̥kʷos (*l is vocalised, your fonts may not place the vocalisation circle correctly.)
But normally we'd expect the reflex of *kʷ in Proto-Germanic to be *ƕ [ʍ] or the Verner variant *gʷ, not *f as seen in English, German and Dutch Wolf. *f can only come from PIE *p. To explain this odd switch between *kʷ and *p some people have proposed a sound law like the following:
*kʷ > *p /R̥_ Or: *kʷ becomes *p after a vocalised resonant. Other such ad hoc explanations could be found, to explain the shift, simply because the environment is quite rare. I for one, would not know any other root with the same environment. Since there's no real way to disprove or approve such a law, we should be careful with accepting it.
But we haven't run out of odd Germanic cognates yet. We also have the word whelp. Arguably this word could be completely unrelated, but the build up is rather similar. Cognates to this word are O.S. hwelp, O.N. hvelpr, Du. welp, Ger. hwelf
A proto-Germanic form of this word would be expected to be *ƕelpaz. The most natural reconstruction of this word in PIE would then be *kʷélbos. Bizarre, I know but the similarity in root is striking. But the Indo-European word has a *b, which immediately makes it suspect. It looks like whelp actually developed in Proto-Germanic rather than Indo-European.
Proto-Germanic form of wolf is *wulfaz. *ƕelpaz looks as if the *w devoiced, and the *f hardened. Besides that the root has e-grade instead of Ø-grade. There really is no proper way to explain such changes, and maybe I'm completely wrong in assuming a relation between the two. But it's worth mentioning in this blog entry about the word wolf. Some propose the word to be onomatopoeic, but the same could be said for the Indo-European word for wolf.
There's one more very odd word ON. Vargr 'Outlaw, Wolf', but ModIc Vargur means 'fox', the Old English word is Warg, and referred to a particularly evil kind of wolf.
This word, if related to the above cited root for wolf, has also gone through some very odd sound changes. The hypothatical proto-Germanic form would be **wargʷaz
First of all, it seems to have a a-grade in the root. Then there's the *g which is a Verner variant of *ƕ. This would then mean at some point the accent would have been on the last syllable at some point. In PIE we'd expect *uolkʷós. Another odd thing is of course the *r rather than the *l. But this shouldn't surprise us too much, r/l switching is a 'common' sporadic shift.
I suggest we now leave the confusing bunch of words that we know as the Germanic wolf-words and focus on some other languages.
First Greek. In Greek we find the word λὐκος. I can see you guys frowning! That's right, the *l and *u switched places, besides that this word reconstucts perfectly fine: PIE *lukʷos. This is not a normal process, not in Greek not in any language. Greek is known for its funky metathesis though (compare γυμνος to 'naked'), but even for Metathesis-happy Greek this metathesis is very funky.
There is a possibility that the word λὐκος comes straight from *ul̥kʷos though. One could assume a Proto-Greek form *ϝλάϙος /wlákʷos/ after which the labial element influenced the ά to colour to ὐ. Not something intirely dissimilar happens to κὐκλος <*kʷékʷlos where the é is influenced by the labial element. We have no indication though, that this happens to *a as well, and even less evidence that this labial-colouring occurred after the appearance of a through the vocalisation of the *l and in fact that seems fairly unlikely to me.
Let's take another word! Latin this time, lupus. Again a *lu- variant rather than *ul-. Interesting back in the days that people suggested that Latin and Greek may have been closely linked, but by now we know this isn't at all the case, which makes it all the more puzzling. I don't know much of Latin but lupus seems to only be able to go back to *lupos. Both the *p variant and the *lu variant in one word! Can it get any crazier? No it can't, and that's why we're now going to discuss Sanskrit!
Sanskrit has वृकः vṛkaḥ. Wow that one goes right back to *ulkʷos! Coincidence? I think not. This is yet another example of the shameless Sanskrit-centricity in almost anything we reconstruct in Indo-European. Something we should discourage. How is it that of the Major branches only this branch truly displays *ulkʷos and therefore, for some reason gets a more important status than the rest? Just because Sanskrit has a lot of archaisms, we can not just assume every word that's different from the rest is an archaism. But okay enough venting of frustrations on Sanskrit-centricity (Hah! how's that for a neologism!).
So, now having discussed several appearances of the word 'wolf' what kind of conclusions are we going to draw from this? In fact I could still continue for example, telling about the seemingly similar root for 'fox' and also 'lynx' which looks like an n-infixed form of lycos. But that would make this entry even bigger than it already is, so I might discuss that some other time.
There's quite a few things we could come up with, and none of them are absolutely convincing. But let me mention just a few ideas worth mentioning
I'm not a great fan of Pokorny, and in my humble opinion, no one should be one. But he does have quite a nice explanation for this word. He proposes that the root *uelkʷ- is somehow related to the verbal root *uelp- 'to yelp', howl etc. This is in fact quite interesting, it would explain the seemingly random *kʷ/p alternations. I don't know of any other roots, but I do know it's not uncommon that *kʷ alternates with *p. Why this is, is absolutely beyond me. But it does give some nice options. Though it leaves the *lu~ul metathesis unexplained.
Other explanations would say that the 'fear' for such wild animals as wolves would give birth to Taboo words, explaining the odd root formations. I find this absolutely implausible. Simply because such metatheses as *lu~ul are much to 'subtile'. When we see words being inverted, as for example in verlan, we see that this is not done per phoneme but often per syllable. It's not very likely that this is truly a taboo formation.
Another theory, which I've been theorising myself (although undoubtedly other people thought of it too) is the following.
Obviously what we know as Proto-Indo-European was not a sole linguistic entity in the area. It's likely there were other languages, and not just other language families, but sister languages of Indo-European itself too. Similar to how Dutch is a sister language to English. If countless languages develop from English and thus becomes the proto-language of a new family tree, while Dutch dies out, it doesn't mean Dutch never existed. Loan words from other such Indo-Europaic languages are not at all unlikely. Nor is it unlikely that such a sister language had a shift like **kʷ>**p. Did we loan the word for wolf from one of these sister languages? We can never be sure.
We could also explain the *lu- variants with this. In one of these sister languages, it's not at all unlikely that a vocalic *l developed to *lu giving a **wlu-like cluster. I find it unlikely any language would retain a horrible cluster like **wl- which would lead to the disappearance of **w.
It is then absolutely possible to imagine sister languages with realistic sound shifts giving the results Indo-European has. But what would be the reason for taking over the word for wolf from these odd people? Why not keep our own?
Another idea might be that there was a substrate language, maybe even something Finno-Ugric, without the phoneme *kʷ, making it *p by sound substitution. Similarly a vocalic *l is easily pronounced *lu when said language has no such thing as vocalic consonants. But once again, why did a substrate language have such great influence on a word that doesn't have little importance within one's basic vocabulary of a prehistoric culture?
All very nice ideas very little conclusions. Feel free to add any idea's, no matter how crazy, I'm interested in what you think.
DISCLAIMER
I'd just like to thank Glen Gordon for making me realise that of course there must have been Indo-Europaic languages around Proto-Indo-European, and not just languages of other languages families. This possibility never occurred to me before, while it's actually so obvious.
So, I had my Hittite 2 test today. It included Hittite, Hieroglyphic Luwian and Lycian. It went fine, but as I was busy working with it, I stumbled upon the last question of the test which went as follows: What can you tell about the genitive in Anatolian languages?
The answer was easy enough, and I'll give it as an introduction that is to come.
Anatolian languages seem to have two separate ways of creating a genitive construction. One is by the typical genitive ending suffixed to the possesor. This is Proto-Anatolian *-as (Hitt. -aš, Luw. -as, Lyc. -eh). This just goes back to the typical Indo-European genitive suffix *-ós (although some might argue it's the other way around).
What is interesting about the *-ós suffix though, is that it is also seen in the Anatolian a-stems (PIE o-stems), while in the later PIE languages we tend to find the pronominal *-eso, or even a *-esio (and maybe *-eio).
The other way of creating a genitive construction is the -ssi- suffix (Luw. -esi, Lyc. -ehi), oddly enough this formation seems to be absent in Hittite, while it is dominant in Hieroglyphic Luwian and Lycian, and the only formation in Cuneiform Luwian. This formation is typically Anatolian and works quite different from what you're used to.
I'll give an example of both the traditional genitive construction, and the Anatolian genitive.
Both forms can be found in a Lycian trilingual on the founding of Xanthos (a Lycian city). The *-as genitive is used as a so called onomastic genitive (a genitive used for names), while the *-ssi- genitive is used for nouns.
ẽke Trm̃isñ χssaϑrapazate Pigesere Katamlah tideimi
'When Pigesere son of Katamla became Xadrape of Lycia....'
The Genitive construction here is Katamlah tideimi. where -ah goes back to *-as (though we would expect **-eh, but there's still many things we don't fully understand about the *a>e shift in Lycian.)
se-t-ahñtãi χñtawatehi χbidẽñnehi sey-ArKKazumahi
'and to be of the king of Kaunos, and of Arkesima'
And there we have, not one, not two, but three -ssi-genitives!
χñtawatehi, χbidẽñnehi, ArKKazumahi. What's interesting (though not particularly visible in this example), about the -ssi- suffix, is that it makes it a new noun, which afterwards, just conjugates like any normal noun. So *-ssi-s is the Nominative, *-ssi-n is the accusative, etc.
So here is approximately where I ended my answer, and, my answer had less cool examples and diachronic notes, but you get the point.
So, now that we have an overview of Anatolian genitives. Let's look at some other Indo-European languages.
First Greek:
Standard for all consonant stems is the -ος genitive. For example:
πατηρ > πατρος
o-stems though, work differently. they end in -ου. For example:
ἱππος > ἱππου
This ending could come from a variety of sources.
Some idea's are: *eso, *eo, *ejo but definitely not *esio. The first *e could also be an *o, and technically the final *o could be an *e if the first letter is an *o. *eo is dismissable if we're to assume it's from an Indo-European origin.
*eso is the most likely, since this is what we see in the pronouns, which in most cases seem to work the same as the o-stems.
Gothic has -is, which should be safely derrived from *eso as well.
But then comes Sanskrit to screw it all up!
Taking the same examples (etymologically) as Greek:
pitā > pituḥ पिता > पितुः
This formation is rather weird, I believe it points to a *-r-s ending, and thus proterodynamic rather than hysterodynamic as seen in Greek. It's not that relevent though, the word for horse (an a-stem (PIE *o-stem)) shows a very odd genitive:
aśvaḥ > aśvasya अश्वः > अश्वस्य
An -asya suffix. This could be from a variety of sources, the likeliest ones are *esio or *osio. It might just be me, but I think the -sy- part looks an awful lot like the Anatolian *-ssi- suffix e-si-o. An *e thematic vowel, *-si- suffix, and an *-o ending. What exactly this *o-ending would be is unclear. Maybe it's an *-e, and then it would be an old vocative suffix. Maybe it's based on the pronominal ending *-eso which might have sounded enough like *-esi to start influencing each other.
When I presented this idea to a fellow student of mine, he pointed out the rather odd dative plural of consonant stems in Greek, which is made with a -σι(ν) suffix as well!
It's hard to explain how a genitive *-si- ended up in a dative plural. But the current explanation isn't too satisfying either. People generally compare it with the Sanskrit locative suffix -su सु/षु. But an ablaut of i/u isn't what we like to see in Indo-European.
So to conclude. There's two types of genitive constructions in Anatolian languages. The true genitive *-as and the genitival stem extension *-ssi-. There are some indications, especially in Sanskrit, that this suffix may have already existed in Indo-European itself, but only ended up being productive as a separate category in the Anatolian languages.
I am quite sure there's other Indo-European languages that have evidence for a -si- genitive like Sanskrit does, maybe Slavic languages? I'm not sure, but if you know, please leave a message.
An interesting feature of the Indo-European verbal (and sometimes nominal) system is so called 'reduplication'.
Reduplication in Indo-European is the repetition of the first consonant plus an extra vowel before the action root.
A famous example is the Indo-European word for 'wheel' one of the few nominal roots with reduplication:
*kʷe-kʷlo-
Note that the reduplication is solely the first consonant, never the cluster. Well never is, of course, a big word. It does seem to happen sometimes.
Typically in Indo-European we reconstruct two types of reduplication, one with a vowel *e after the reduplicated consonant, one with the vowel *i. Obviously we like to reconstruct some kind of earlier **ə that gave both reflexes, but in the latest phase of PIE these two were definitely there.
Are these reconstructions really that natural though? Maybe not. Many languages seem to indicate a different system. For once Sanskrit was not authoritative in the opinions of the earlier Indo-Europeanists (typically, when there's doubt, the Sanskrit form is considered correct). No, this time it's Greek. Greek knows solely two types of reduplication. i-reduplication in presents (and very rare aorists), and e-reduplication in the perfect.
When we have a look at three other languages with attest reduplication, we get a very different view though. The languages I will discuss are Gothic, Sanskrit and Tocharian B. I've been told that Latin also displays some non-standard reduplication, but since my Latin isn't much better than 'dulce et decorum est pro patria mori', I'll leave that languages for people with more expertise.
First, let's talk about Gothic. Gothic is a nice language, for it displays reduplication often, and consistently throughout it's full history (which is awefully short, but still), unlike any other Germanic language. There's some proof of reduplication in other languages (I believe an odd Old-English form for 'rowed' was 'roard' (though spelling is probably wrong)). The other for is what we see in the preterite of do did < *deda. But the *deda discussion is a different one all together where no consensus has been found yet.
Gothic reduplication, occurs in the preterite. This is exactly the way we like it in Indo-European reconstructions. There is one main issue though. The vowel use for reduplication is <ai> which phonetically is [ɛ] (some people are pro diphthong, especially in the long variant, I disagree, it has to be a monophthong).
This is not the vowel we expect in reduplication. PIE *e consistently became Gothic i in all contexts. Later though i was changed to ai in several contexts (this all happened before our first Gothic textual attestation).
*i > ai/_r,h,ƕ This is linguist speak for 'e became <ai> before r,h and ƕ (Similarly als *u > au/_r,h,ƕ for you symmetry lovers).
Not all reduplication verbs start with r,h or ƕ though. In fact I can't think of any. So a verb like letan you'd expect to have a preterite **lilot. What we actually have is lailot. Some people have desperately tried to explain this to say that the vocalism is analogy from the reduplicating verbs that do start with r,h and ƕ. Personally I do not find this very convincing. Nevertheless, it's very hard to have any <ai> in Gothic outside of the above mentioned position. Ony if it were the long variant of <ai> (not visible in writing), we would be able to have that vowel outside of that position.
If <ai> in reduplicating verbs is not from earlier i, we would expect it to be from an earlier *ai < PIE *oi, which is not at all convincing either. Because of this (and several other reasons), I think there might be reason to assume that reduplication in Gothic is not a direct descendant of Indo-European reduplication (this assumption is going to academically kill me :P),
If it is old though, we have either a very odd Indo-European reduplication, or a very odd development in Gothic. You choose ;).
Now, let us discuss Sanskrit. Sanskrit uses reduplication in much the same contexts as Greek. It's used in some presents/aorists (which, I believe, use exclusively the i-vocalism). Besides that the i-vocalic reduplication is used in desideratives and intensives (some indication of this, is also, and solely available in Greek (even more proof for my pro-Graeco-Iranian theory :P)). In the perfect though, we see a different situation all together. We don't see the velar-palatalising a that we would expect as a reflex of PIE *e. No, it reduplicates the sonorant semi-vowel or vowel of the root in the zero grade.
What does that mean? Here let me show you:
Stem: puṣ- present: puṣyati perfect: pupoṣa (< Vedic pupauṣa). So it actually reduplications the u. This is odd in many ways, especially because the sound u is not actually immediately following the first consonant when the verb is in the perfect. In the perfect plural it is though, so likely it got it from there.
This reduplication is also seen when the semi-vocalic consonant is an i. Also the a is reduplicated when it's in the zero grade of the root, or if the second consonant is a vocalic r. But to be honest, this might be the Indo-European *e, especially since it palatalises the reduplicated velars.
Note that this semi-vocalic root vowel even crosses cluster boundaries. So the root śru- gives śuśro-.
Another distinctive feature of Sanskrit reduplication is the way the consonants are treated. When a voiceless non-aspirate is reduplicated, the same consonant remains. An example was given above.
If a voiced aspirate and h < *gh, is aspirated, a voiced non-aspirate. For example
dhāv- > dadhāv
This is not that different from Greek, it's Grassmann's law. Only Greek has voiceless aspirates, and they thus give the voiceless non-aspirate. (τιθημι).
But in some roots, and this is really weird, a voiceless consonant reduplicated sometimes gives the most bizarre reflexes. One that comes to mind right now is the verb pibati 'he drinks' with the root pā-. That's right, the root shortened it's ā to a, then the root consonant became b while the reduplicative consonant remains p. Absolutely insane, but very real nevertheless. This is but a minority though. Most voiced reduplicated consonant remain voiced.
This is, once again, a very different form of reduplication than the one we see in Greek. I've been told that something similar happens in the few Latin verbs that have reduplication. But someone else is going to have to give examples.
Now we get to the last, and really crazy part. That's right, Tocharian B. Tocharian B (and probably also Tocharian A :P) is always crazy. It fails to conform to any type of normal behaviour. I love it.
Tocharian B has reduplication in only one class. The Preterite Participle.
The most normal Preterite Participle is perfectly normal. CäCaw (a<accented ä). Example:lätaw from the root lät-
The second group is C1eC1aw (a from accented ä). Example śeśaw. Though this looks like a perfectly normal reduplication, this is already bizarre. Because TochB e does not come from IE *e, rather it comes from IE *o. This o is also nowhere to be found in the root. śä-
The third group is C1eC1aC2äw (a from accented ä), For example lyelyakäw from root läk-
Next group is C1eC1éC2äw Example: teteku (u = äw) from täk-, At last here we find a decent reduplication of the root vowel.
(j)VCäw where V is accented. For example: yáku root: yäk-
C1aC1áC2äw for example: lalālu (ā = accented a) from root: la-. TochB. a comes from Indo-European *h2e.
The next class is rather normal (no reduplication!) CäCáu for example: krämpau root: krämp-
The before last class is C1aC1áC2au for example kakārpau root: korp-
The last class I don't have any example of, but luckily isn't reduplicating either. áCau.
So in Tocharian we have reduplication with o, or h2e. And not once with *e or *i. Nor is the root semi-vowel reduplicated as we'd hope to see considering Sanskrit.
So there we have it! Three completely different reduplication systems, which are opposed directly against the Greek and reconstructed Indo-European form. Why we are still reconstructing the reduplication the way we are is actually a bit mysterious. It's the easiest, that's for sure. All these systems might have also undergone some analogy, especially Sanskrit and Gothic.
This leaves us with Tocharian B though, which went about being it's own crazy self (he likes to do that). Now we might want to mention that Tocharian split of from the Indo-European family well before the rest of the languages started branching off (but well after Anatolian branched off). Maybe this was indeed the crazy reduplication system PIE used to have, but then all reduplication was reduced to a schwa.
One thing is sure, the last word has definitely not been spoken about Indo-European reduplication, it's no where near as straight forward as some people like to think (Though in all honesty, to keep your sanity, I can imagine you'd rather think it's that simple than look at Tocharian; What can I say? I'm a linguistic masochist :P).
Okay, Greek isn't really a satem language, but it is a good introduction to this post. First, let me point out that I have no idea whether people have researched the theory I'm about to propose before. I'm guessing that I'm in all likelihood not the first one to notice this.
Traditionally we like to split up Indo-European into two main dialectal groups. The Central and Peripheral group. The central group are the so called satem languages, due to the result of the velars in these language. The Peripheral group are the centum languages.
For more information on what actually distinguishes satem from centum read my older blog entry: One Hundred
Now then, let's get to the point. It is true that there is definitely a tendency of Centum languages to be more similar to each other, and similarly Satem languages to be more similar. Nevertheless, I don't believe that the Satem-Centum differentiation was a big enough shift to completely disconnect the two groups as two main dialects not able to influence each other any more.
I even believe that Greek, a Centum language, shares far more feature-isoglosses with the Indo-Iranian languages than it distinguishes itself from those. Therefore, I would like to propose a Graeco-Iranian dialect, rather than an unlikely incredibly similar development of Greek next to Sanskrit without any influence on each other.
Note, that I will not go as far as saying that these languages should be considered a 'branch' of the Indo-European language family. Family trees are starting to become painfully outdated, and are only an easy abstraction when talking about very general groups.
Let me explain why I believe there was a Graeco-Iranian (henceforth G-I) Indo-European dialect.
1. Productive reduplicated perfect
Traditional Indo-European linguistics tells us that the standard form of creating a perfect was reduplication of the first consonant and a o-grade in the root.
The perfect stem of *gen- 'to be born' thus be comes: *ge-gon-. This is terribly biassed towards Greek and Sanskrit (which is a very typical fault in traditional Indo-European reconstructions I must add). Greek and Sanskrit clearly show this system as being the productive one. The most well known perfect that doesn't have reduplication is *ueid- 'to see'
which in the perfect is *uoid- 'to know' (Compare Dutch 'weten', and Gothic 'waitan').
In itself it's very mysterious both Greek and Sanskrit share this as a non-reduplicated perfect, although there's next to no other verbs that do this.
So from a Sanskrit-Greek bias we conclude reduplicated perfect is the Proto-Indo-European form.
If we look at other languages though, we get a very different view of how the perfect is formed. First let's point out that Anatolian has no indication that there was reduplication at that point in Indo-European when the Anatolian languages split of. There clearly was a distribution of verbs with either e or o vocalism though. But this is of lesser importance, we can conclude that reduplication was a later development. Tocharian (one of the other early differentiating languages in Indo-European) there's already proof of reduplication, but solely in the Perfect Participle.
Now we get to Germanic, which will be one of the important branches that will support my theory of a close relation between Greek and Indo-Iranian, and the terrible bias towards those to languages in our reconstructions.
Germanic as we know it today has next to no proof of reduplication. Except for of course, the preterite of do: did <*deda. Old English apparently has some words which might be traced to reduplication, but it's hard to tell.
What Germanic does have though, is o-vocalism in the preterite stems of the strong verbs. almost strong verbs have o-vocalism except for the classes I'll get to later. Know that Indo-European *e often becomes Germanic i, and *o always becomes a.
Examples of o-vocalism in English:
give ~ gave
see ~saw
bear ~bore
lie ~ lay
Gothic though, as the only Germanic language retains a productive category of a reduplicating preterite.
For example
saian ~ saiso 'sow, sowed'
waian ~ waiwo 'blow, blew'
letan ~ lailot 'let, let'
Nevertheless, this is only one of the seven classes of preterites. But etymologically this can be brought back two two classes. Non-reduplicating perfects and reduplicating perfects. It's not unlikely for a language to get rid of its reduplication. But retaining two productive systems is highly unlikely. In Gothic the reduplicating class was a minority class, but definitely an existing one.
I therefore believe that Indo-European had two types of perfects reduplicating perfects, and non-reduplicating perfects. And I tend to believe that the non-reduplicating perfect was the normal variant, whereas Graeco-Iranian had made the reduplicating perfects the productive class.
2. Grassmann's law
Both Greek and Indo-Iranian languages undergo the so called, Grassmann's law. This law says that no two aspirated consonants can follow each other. The first consonant will lose its aspiration. These laws only take place after Greek devoices its aspirates. therfore we have trikhos from earlier thrikhos. And we have Sanskrit Bodhati from erlier bhodhati.
That this law took place, and spread itself over both the Indo-Iranian group and the Greek language is not unlikely. I do not think that the initial differentiation of Greek's aspirate devoicing is enough for the two very similar dialects to lose touch. I thus believe that Grassmann's law was not a law that developed in two different branches, but rather its 'wave' spread itself over both dialectal regions, simply having a different effect due to their initial differentiations.
For more information about this 'wave' theory of dialectology and linguistic differentiation I refer you to Language waves and the satem innovation in PIE.
This is it for now. I had a lot more in my head, and it has currently flow out of my head (remind me to make notes from now on!). There is some stuff on the pronominal system in Greek and Sanskrit that makes me go Hmmmmm, but it is difficult to prove. I also believe that maybe the 8-case system as seen in Sanskrit is a Graeco-Iranian innovation. Although Classical Greek only has 4 cases (and a vocative), there's clear indications of older cases. But these are all speculations. I will write more once I get to it. Personally though, I think the productivity of the reduplicating perfect is a very strong indication of a seemingly close relationship.
I come up with the worst puns for my article titles. But okay, let's talk about the the Tocharian verbal root yäm-.
As I was learning my Sanskrit vocabulary I came across a very odd word: dā- दा 'to give'. A perfectly normal Indo-european verbal root related to the Greek root δω- 'to give'. But when we have a look at the present of this verb, is where it gets odd. The present in Greek is δίδωμι, typical reduplicated present athematic verb. No such luck with Sanskrit though, Sanskrit actually uses a completely different root in the present, namely yam- यम् giving a first person present yacchāmi यच्छामि.
This is what we call 'suppletion' when there's multiple roots in one paradigm. Especially in common words it's to be expected to see a lot of suppletion. Also, due to Sanskrit's extremely big vocabulary, it tends to sprout a lot of suppletion. Usualy though, this suppletion is through loans from the Dravidic language, as I also expected to be the case with yam-. But then suddenly, in a moment of clarity, I realised that yam- has a cognate in Tocharian, namely yäm- `achieve, obtain; reach'. I realise this actually means the exact opposite. But it's still the concept of physical transference, and it's hard to believe these words are not related.
The Indo-European root would then be *iem-. Now, the Tocharian languages (A and B) are known for their extensive suppletion. So I was secretly hoping that Tocharian would display the same suppletion as Sanskrit, so we could conclude some kind of ancient Indo-European suppletion, but no such luck. In fact, the Tocharian B verb doesn't have any suppletion at all!
Adams' Tocharian B dictionary:
Paradigm: G Ps. X /yänmāsk'ä/e-/ [A yänmāskau, -, yänmāṣṣäṃ//; APart. yänmāṣṣeñca `achiever'; MPPart. yänmāskemane; Ger. yänmāṣṣälye*]; 1Ko. I /yonm-/ [A -, -, yonmäṃ//]; 2Ko. VI /yänmā-/ [A -, -, yanmaṃ//-, yänmācer, -; MP //-, -, yänmāntär; AOpt. -, -, yanmoy//; Inf. yänmātsi]; Pt. IIIa /yonm-/ [A yonmwa, yonmasta, yonmasa//]; PP /yeyänmu-/;
K Ko. IX /yänmäsk'ä/e-/ [Inf. yanmäs(t)si]
Just a 'normal' verb. Although actually, it's quite a disturbing paradigm. n-infixed ske/o-suffixed present. With accented -a- suffix in between the root and the ske/o-suffix. in the present. Lovely. But in Tocharian it's definitely not unheard of.What is crazy though, is the reduplication in the Past Participle. That's a reduplication with Tocharian B e, this might not look crazy, but believe me it is. TochB. e comes from PIE *o. We've seen *e in reduplication, *i in reduplication, but *o?! And just to mess with us, it also adds 'a' (<h2, h2e) reduplication in some other paradigms.
Makes you wonder what the hell those Tocharians were on.
Now, I figured for completeness I'd add the paradigm of 'to give' in Tocharian B, just for completeness, though I can assure you you're not going to find the preferred *deh3 in there either.
Paradigm: Ps. IXa /āisk'ä/e-/ [A aiskau, -, aiṣṣäṃ//aiskem, aiścer, aiskeṃ; AImpf. -, -, aiṣṣi// -, -, aiṣṣiyeṃ; APart. aiṣṣeñca (see also aiṣṣeñcaññe below); MPPart. aiskemane; Ger. aiṣṣälle]; Ko. I /āi-/ [A āyu, ait, aiṃ// -, -, aiṃ; AOpt. -, -, āyi//; MPOpt. -, -, ayītär//; Inf. aitsi; Ger. aille]; Ipv. /pete-/ [sg. pete; pl. petso ˜ petes]; Pt. I(sg.)/III(pl.) /wäsā-/ [A wsāwa, wsāsta, wasa//wasam, -, wsar ˜ wsare]; PP /āyu-/
ai- joy, what a great root that looks nothing like what we'd like to see. It is sometimes connected to the Greek αἴνυμαι 'to take, seize'.
But luckily this paradigms has some nice suppletion. A Preterite with the root wäs- ! Go make up your own etymology for that, Adams' best guesses are as good as anyone's.
And besides that, there's a very interesting Imperative. pete. Tocharian had a prefix p- or pä- for imperatives (not pe-!) which would be prefixed to the verbal root. How exactly the prefix came to be pe- is uncertain, maybe it was affected by the following root te-. te- is interesting though, it looks like our long lost *deh3. It doesn't go too well though, we're not to sure what to do with PIE *ō's in Tocharian (or *oH or *eh3 that merge with *ō), it tends to become u, but sometimes it becomes e. taking a reconstruction of the root like *dh3e- would give us a much more safe reconstruction as te-. but there's not a clear reason to assume this.
Although not a guide to the Tocharian verbal system, I hope you've gotten an idea how incredibly crazy, and thus, incredibly interesting it really is.
What's that? Want to check out Adams' Tocharian B dictionary yourself? Go to IEED and click the link. It might take a while, the server backing the database is terrible, but after a while it starts working.
Take note though, Adams is notoriously sloppy, and his proposed etymologies are more often than note disposable. He often has rather irregular sound laws. Besides that he doesn't quote textual sources too well every time, he does quote them quite consistently though, which I guess is a plus. The translations he provides can also be rather sloppy.
Some resources:
Indo-European Etymological Dictionary - Nice site with a lot of dictionaries, including the digitalisation of the only Tocharian dictionary out there. But a server you'll instantly come to hate.
TITUS Project - A fantastic resource of an enormous amount of digitalised and transliterated Tocharian texts, both Tocharian A and B. Be sure to check this one out!
I'll interrupt my little series on Tocharian B (which I will continue) to talk a bit more about my theory that maybe /j/ in Proto-Greek was voiceless rather than voiced. I came to this conclusion completely without the intention to do so when I started to write an article to prove that /tt/ was in not a long consonant.
On review, /tt/ definitely is a long consonant, reading the Iliad at university made me realise without this assumption a lot of the poetic meter wouldn't work.
Nevertheless, I hypothesised that the /j/ in combination with <κ> <γ> and <χ> must mean it's voiceless to create a voiceless geminate like <ττ>. There's also a clear indication that <ττ> probably was closer to a [c:] rather than a [t:] due to the later development into <σσ>.
Altogether this would mean /j/ is in fact voiceless and would have been closer to the sound [ç] rather than [j]. Now there is actually some other proof that this /j/ was voiceless, at least in several other environments. I came to notice this as we were discussing the initial consonantal PIE *i- reflexes in Greek in my class of Historical Grammar of Greek.
Oddly enough, PIE *i- seems to have two reflexes, namely /h/ or 'spiritus asper' and <ζ>. /h/ is not an unlikely development you'd expect from a word initial /ç/, it's similar to Proto Germanic where *x became h in initial position.
The reflex <ζ> is more likely if you assume the phonetic value of /j/. It's just like how in French the initial i from Latin came to be pronounced as [dʒ] and later just [ʒ].
So why were there two reflexes for PIE *i-?
There's two theories which oddly enough, contradict themselves.
Some say /h/ comes from *Hi- with a laryngeal before i rather than a plain i. And that the <ζ> is from a plain *i-.
Others say that /h/ comes from *i- and <ζ> comes from *Hi-!
Isn't that wonderful? At my university the first theory is supported due to a single example in Sanskrit that seems to indicate this. Greek knows a word ὑσμίνης /husmínês/ 'combat' which in some terribly convoluted way seems to be related to the Sanskrit root yudh- युध् 'to battle'.
This doesn't tell us anything yet, but there is a certain compound verb in Sanskrit that seems to indicate that there's an initial laryngeal. this word is amitrāyudh- अमित्रायुध् 'to fight against those who are not friends'. This word consist of the elements: a- 'not' mitra- 'friend' yudh- 'to battle'. As you can see, for a mysterious reason the a in mitra- gets stretched. This is not a normal reaction for the noun in this position and, can only be explained by assuming there was a laryngeal in between mitra- and yudh- *mitraHyudh-. a laryngeal behind a vowel gives a long ā in Sanskrit, and this would prove that there was a laryngeal in front the indo-european stem *Hiudʰ.
There's one big problem with this though, we don't have any counter examples to prove that there wasn't an *H- in front of the *i that gives the Greek reflex <ζ>. In other words, we have proof that the /h/ reflex comes from *Hi- but no proof that <ζ> comes from *i-. With this hole it's not yet safe to assume anything.
The other theory is not so much based on proof, but rather on phonological reasoning. And Indo-European *H could be vocalised, in other words, it could be the nucleus of a syllable. It's extremely difficult to make a voiceless consonant the nucleus of a syllable, and therefore we often assume that the laryngeals were voiced consonants, which ones exactly, we don't know.
Now if you think about it this way, it becomes increasingly hard to believe that a sequence of two voiced consonants became a voiceless /h/. Therefore, one might want to assume this sequence became <ζ> and simple *i- became /h/ instead. But the first theory has more indirect proof than this one, as far a I know.
The latter theory does fit in better with my /j/ is a [ç] hypothesis. If /j/ were voiceless [ç], it would be quite likely for it to become a voiceless [h] in initial position. Also the other way around though, it's not completely unlikely to assume that if for some reason we're dealing with a voiceless laryngeal, a voiceless /j/ wouldn't be bothered to become an /h/ in initial position.
Which explanation is the right one, or if either of the two are wrong is hard to tell, one thing is for sure, we're not yet close to finding a definite answer to our questions about the odd reflexes in Greek of the PIE *i-.
NOTE: Halfway through my article, I just found an enormous gap in my reasoning which I'm going to pick out myself, before someone else does. I'm not sure if we're certain that the initial /h/ was actually voiceless! for all we know it was a voiced glottal fricative [ɦ]. The only thing that would be possible to assume as proof for a voiceless /h/ is what we see when a unvoiced unaspirated stop comes before an /h/ initial word. Namely it becomes a unvoiced aspirated stop C + h > Cʰ, while if /h/ were voiced you'd expect a C[+voice]ʰ. Of course there's a gap in this reasoning, because there's no such thing as a voiced aspirated stop in Greek, nevertheless you'd sooner expect the combination to form a voiced stop, rather than a voiceless aspirated stop.
By far the most frustrating part of Indo-European linguistics is the Aorist and presentive suffixes, which are actually two sides of the same problem. Let me explain:
The Aorist is a sort of past tense, and mostly attested in Greek and Sanskrit, I'm not actually sure of other languages, at least Anatolian and Germanic language have no indication at all that a thing like an Aorist may have existed.
Semantically, over at my university, Leiden, they seem to be convinced that the Aorist gave more 'focus' than the normal past tense also known as the imperfect. I'm not sure about Sanskrit. post-Vedic Sanskrit has lost most uses of the Aorist, and it is mostly seen used as a prohibitive (in combination with a negating particle).
The Aorist's existence is really only a fact in the Greek branch and the Indo-Iranian. Morphologically it can be distinguished from the imperfect due to either the presence of an -s- the so called Sigmatic Aorist, or by the absense of present markers the so called Normal Aorist.
The Sigmatic Aorist is fine, I can easily understand it and it is the Aorist that became particularly productive in both Greek and Sanskrit.
But the 'Normal' Aorist, is incredibly puzzling. Before going into the Normal Aorist though, I'll have to explain what said present markers are.
Present markers are augmentations to the verbal root, usually suffixes, but also prefixes (if you want to call reduplication a prefix) and infixes exist. In some languages these markers have very clear functions, in other ones less so, and in most languages the marker sometimes seems to have a function, and sometimes it just doesn't.
the *-ske/o suffix for example has quite a clear 'iterative/intensive' use in Hittite, though there are plenty of examples in which such a meaning can't be recovered, or where there is no instance of a suffixless form of that verb.
In Tocharian B it is either just a 'usual' present tense marker or it can also make some causatives. And in Greek it is only a, what we call, 'present strengthening. Which means both Present and Imperfect have the suffix, while other tenses (Perfect, Aorist and Future) do not with no apparent change in semantics.
Obviously, linguist don't like to have a legion of suffixes which have no apparent meaning and are attached the root without a clear aim (except for marking the present). Besides that it feels unnatural to make the present more 'marked' than a derived tense (and linguistic 'universals' will agree with me).
To explain this terrible distribution of suffixes and aorists I've found a theory which may or may not work. But, it's purely hypothetical and can't really be proven wrong, it's the only way I can make them work in my head though.
I suggest that originally there was a plain verbal root. This root could be declined in several tenses. Oblique (present), Aorist and Perfect. (I highly doubt the existence of an Indo-European Imperfect, similar to the Plusquamperfect in Greek). Then there was also the *h1e- augment prefix to mark past tense, and the secondary endings to mark the past tense, this could then be used on the Oblique, Aorist and maybe Perfect (though indications of the last are weak).
Aorist would always have secondary endings, even when unaugmented. Of course, the unsuffixed Oblique would be absolutely homophonous to the Aorist in the past tense. In other words, using the 'Aorist' to give focus was only possible in the present tense.
Besides all this, the Oblique could be augmented with the present suffixes, to alter the meaning, creating causatives, iteratives etc. These suffixes were not applicable on the Aorist, in other words, suffixed verbs could not be focused like the Oblique could.
This whole story has proven to be quite confusing, so let me sum up my theory once more.
Three 'tenses': Oblique, Aorist, Perfect
Oblique takes suffixes to augment the meaning.
Aorist doesn't but gives more focus that unsuffixed Oblique.
Perfect doesn't either.
Now the later development gets a bit odd. Maybe something like this happened: As mentioned earlier, Aorist and Oblique were indistinguishable from each when given the augment to indicated the past tense. Maybe the Proto-Indo-Europeans felt that they also needed to be able to denote focus in the past tense, and would thus carry their suffixes into the root form, meaningless, so they'd be distinguished from the Aorist in the past tense. This at least, seems what has happened in Greek.
The only mystery now left is why the hell did the Aorist almost completely disappear in the present tense? Especially in Greek, presentive use of the Aorist can only be achieved with participles + to be. MYSTERYYYYY