Sanskrit isn't Indo-European?

I just ran across an article—maybe excerpts from a vanished article—that suggests that it is colonial to claim Sanskrit as Indo-European. Please note, I may refer to things with rather strong language, but pseudoscience deserves no less.

Some further reading (here and here) makes me think that this is some sort of reaction against perceived erroneous history. The gist of it is this: the Brits when colonizing India saw that Sanskrit shared deep similarities to Ancient Greek and Latin. If that is the case then there must be some common language. Since white people—according to the grossly misinformed thought of the time—were superior, and people in India come in a variety of skin colors, then there must have been white invaders who brought Sanskrit to India. Apparently the colonizing Brits couldn't imagine that maybe, just maybe, the initial Indo-European population had skin darker than theirs. Pity their lack of imagination.

The problem with the old notions is that they are junk. This junk in turn, rightfully, raises a horrified reaction among Indians. So far so good. One should be outraged by bullshit pseudoscience. The problem is when one reacts to junk by inventing more junk. The pseudoscience starts off with a howler.

For instance, in Hungary, there is a growing body of scholars who are extremely uncomfortable and dissatisfied with the manner in which Hungarian was excluded from the Indo-European framework.

I feel bad for the Hungarians who want in on the IE, but that's how science goes. It doesn't always come back with the answers you want. Tough. Believe what you like, the science is the science.

there are some Indian language scholars who have suggested that a computerized analysis of Sanskrit and Latin lexicons might yield a far more limited overlap than would be rationally implied by the “Indo-European” classification.

Of the tiny bits I know about  Sanskrit, I know that counting to ten is an awful lot like modern Persian and telling your name is an awful lot like Latin. This is daily use, low rate of change stuff. I also wouldn't expect there to be massive overlap between Latin and Sanskrit—particularly when getting into descriptions of local flora and fauna. This book has a whole chapter about non-IE words borrowed by Sanskrit. Fancy that, a language that borrows words.

…Building primitive lexicons that show similar roots for certain common words can hardly be an adequate basis of linguistic classification.

Actually this is what I'd expect. Quick, say the past tense for see and dive. What were your answers? Saw and dived are the correct answers as of this writing, though dove is making inroads. Highlight the blank if you are curious. One is not being changed, the other is. Care to guess which? If you said that see is maintaining its historic form, you'd be right. Why? Because it's a basic word in heavy use. Dive? Not so much.

Moreover, it [IE theory] has strengthened the now increasingly untenable view that there is no continuity between the Indo-Saraswati Harappan civilization and Vedic civilization, and that India’s languages (both in the oral and written forms) must have been brought to India by more “civilized” outsiders.

The same way that English sprung up in England without any help from continental invaders? While I have no doubt that language displacement can also disrupt cultural transmission, I don't know that language displacement necessarily causes cultural discontinuity. One (cultural discontinuity) has happened without the other (language displacement) in Greece. But I'm not an archaeologist. Until you've got better evidence, stick with what the data tells you. The data, so far as I can see, tells me that there was some sort of linguistic invasion. These things happen, otherwise I might be writing this in a Celtic language.

In this entire body of work stretching, from Sakatayana to Panini, there is virtually nothing to link Sanskrit to any European influence.

No shit. I wouldn't want to put words in Sir William Jones's mouth, but I'd venture to say he didn't know about the connection and was surprised at the similarity too. Oh, why don't I just let him speak for himself.

…yet [Sanskrit] bearing to both of them a stronger affinity, both in the roots of verbs and the forms of grammar, than could possibly have been produced by accident…

I hate it when that happens. I also hate this too.

both Sanskrit and Tamil are syllabic languages and both treat consonants and vowels very similarly.

All languages have syllables and treat consonants and vowels very similarly, when viewed through the lens of being syllabic. Well, except for maybe a pesky example or two in the Caucasus mountains. Nice try though.

From the point of view of classifying languages based on the organizational principles that govern their written scripts no logic would permit the Sanskrit-derived North Indian langauges to be placed in the same language group as the European languages.

Writing systems don't count at all when comparing languages. Just the spoken end. I get the sensation that the author isn't a linguist.

Writing in Language in India (9, Jan, 2002), G. Sankaranarayanan observes how repeating words and forms is a significant feature that extends across the Indian subcontinent and includes not only the Sanskrit and Tamil derivatives but also Munda and languages from the Tibetan-Burmese group.

It's called sprachbund. It happens all over. The interesting part here—to me anyway—is that the Indian sprachbund includes non-related languages. Nothing to see here.

Note too that Indic languages permit the dropping of pronouns (which become implied). …would be impermissable in English.

Dropping pronouns—even obviously assumed content words—is fine by Latin. No one disputes that Latin is IE. Spanish also thinks dropping subject pronouns is fine. Dammit. I hate it when I pick bad examples like English to almost but not quite make my point. Moving along, our pseudoscientist goes on to talk about word order.

In this respect, Indian languages are similar to each other [in regards to a more free word order] but not to less flexible “Indo-European” languages like English. On the other hand, Russian and Czech (like Hungarian) [no cheating, you need to compare IE to IE] do not require a fixed or default word order.

Oops. I think you're hurting your own argument there.

In conclusion, it might be stated that the present scheme of bifurcating Indian langauges into the “Indo-European” and “Dravidian” scheme is unsatisfactory in many ways.

For whom?

…it has also precluded comprehensive comparitive studies between these Indic languages and other Indic langauges such as the Munda or those from the Tibetan-Burmese stream.

Who is being stopped? These studies might actually be fruitful, but they only count when peer-reviewed.

Sh Thadani [the author, has] a Post-Graduate degree in Computer Science from Yale…

Aha. I was right. Not a linguist. Look, I'm not claiming authority here, but if you're going to write a "research" document it may help to have some formal training in the subject at hand. I don't trust my plumber to fix my car, why does a computer science guy get to look like a linguist?

Anyway. This has put some fire under me to learn Sanskrit. I think I'll put some effort to it over the summer. Well, once I've got True Story read. 

3io and 4th conjugation

I went to a seminar at the University of Chicago yesterday, and, to be quite frank, the lecturer, Benjamin Fortson, was talking about things I didn't know much about despite the topic of the talk being the Latin passive infinitive in -(r)ier.

I did get some really cool take away from it, so it wasn't a waste of time. For the folks who are deep into IE historical linguistics, this is not news. For the rest of it is. I'd cite sources, but Prof. Fortson treated it like it was no big deal, so I guess he's my source for everything that follows.

I'm ditching the macrons and IPA to keep this visible for everyone. Here's where the 3io comes from. First the ugly chart.

verbs in -ye/o
⬋	⬊
“light” monosyllables	“heavy” monosyllables and polysyllables
*kap-yeo	*aud-yeo
⬇	*aud-iyeo
capio	audio
this conjugation's i is short	this conjugation's i is long

Ok, so here it is in steps.

In the group of -ye/o verbs, one of two things can happen. If your root is a "light" monosyllable, you may go down the left side—but you aren't required to. (Light monosyllables: a short vowel followed by only one consonant, which is to say the syllable is open when combined with an ending.)

1. Standard -ye/o verb.
2. -y- contracts to short i.
3. 3io conjugation: now you know why the i's in the present tense are all short.

In the other group, some "light" monosyllables, "heavy" monosyllables and polysyllabic groups, the sequence is this:

1. Standard -ye/o verb.
2. An epenthetic i sneaks in.
3. -iy- contracts to long i.
4. 4th conjugation 

Cool. So far so good. As you can see, the left side works for capio. If you start with *fakyeo, you've got the light monosyllable *fak-, so you're ok to get to facio. Same with *gradyeo. Though it may have been nasty deponent way back when, so let's pretend what I'm showing is the case since the rule applies deponent or not. *grad- is light, so the left side is ok. You can wind up at gradior as a 3io using this scheme.

Obviously being a light monosyllable is no obstacle to going down the right side. Check out *venyeo. Follow the steps, and you get to venio. Heavy monosyllables, like *audyeo go down the right to yeild audio. Polyllable roots go down the right too. So with *sepelyeo, you get sepelio.

So far, so good.

But what about *egradyeo? Sure, it's related to *gradyeo by compounding, but its root is a polysyllable. So 4th conjugation right? What? Cicero says 3io? Sure enough:

…si ex Syria egredi atque irrumpere in meam provinciam conarentur…
                                                      —ad familiares 15.2

It has a typical 3rd conjugation infinitive form. What happened? Rules get broken? Well, first let me show you something else:

foras egrediri video lenonem Lycum.
                                                      —Plautus, Poenulus 742

Hey, wait a second. That egrediri looks like a 4th conjugation form. Well, Plautus must think it is. So somewhere between Plautus and Cicero, egrediri moved from 4th to 3io by analogy. 

Anyway. Here's the cool thing you can take away from this mess: In 3io verbs, CVCio, the V is short. 

The other cool thing I learned is that fio takes over the passive of facio by suppletion. I had always wondered why fio seemed like some awful anti-deponent.

-que and καί

Ok, I was reading a book about Latin's history the other day [well, not really, this is lightly-revised, reblogged material]. I know—just what you do for fun too. This entry may give some insight into how Latin and Ancient Greek relate to each other. And why I find this stuff so interesting.

In any history of Latin in its earliest days, the ancient languages of Italy come up. So far as linguists can tell, there are two main groups. The first group is the Etruscan language. It is a language that no one fully knows. It is limited to a handful of inscriptions, many of which are funerary. Etruscan played an important role in the development of Latin—particularly the Latin alphabet—though it lent many words to Latin as well.

The other group of languages is the Italic branch of the Indo-European languages. Latin is one of these. The others include Oscan, Faliscan and Venetic, although Venetic may not be Italic. It's always hard to be certain about things this far in the past, but to my inexpert eyes Venetic seems similar enough to Latin. We know about Venetic—hold your breath for it—through inscriptions.

The thing that struck me was from an inscription from Este. Because I have a font that can show you the Venetic alphabet, I'll take a stab at transliterating it back into Venetic. I promise zero actual accuracy on my transcription, though I've made efforts at it. The inscription starts on the right and goes to the left.



Now on to the serious part. It is transliterated, from left to right, as:

mego donasto sainatei reitiai egeotora aimoi ke louderobos

Which corresponds to the Latin:

me donavit sanatrici reitiae egetora aemo liberisque

And in English:

Egetora gave me to Reitia the healer for Aemus and his children.

There's all sorts of fascinating stuff here, but the one that struck me was the use of ke. Why? Because Latin has a very similar word it can use to mean "and." The word is "-que." I put a hypen in front of it to show you that in Latin it attaches itself to the end of a word instead of being its own independent word. As you can see, "ke" is its own word. Obviously these words are related. Here's my guess at how.

At one point, there was a word *ke either in Latin or Italic—the star is there to show you that it is not an actual word, but a hypothetical form that ought to have existed. You should not be surprised to find out that a language can go from a "qu" sound to a "k" sound. In English we have King and QUeen. More importantly in Latin we have "quomodo"—with the qu sound—which becomes the Spanish word "como"—with the k sound. So it is very possible to switch between those sounds. So *ke become ke in Venetic and -que in Latin.

I've also noticed a bit more going on too. Maybe I'll blog again about this.

I mention all of this because there is another Indo-European language that uses a very similar sounding word to mean "and." It is Greek, and the word is και, which you might write as "kai" with our alphabet. This word is the same in ancient and modern Greek—so far as my barbarian eyes can see. Again, it is easy enough to get from an "e" (like the "ay" in way) to an "ai" sound (like the word eye). Latin itself made that change, otherwise we would have "praedictions" and "praedispositions." But we don't, because Latin went from an "ai" sound to an "e" sound for the vowel combination of "ae."

So *ke made its way from whatever ancient language—probably Proto Indo-European—into Greek and Latin. This sort of analysis of words can help students made connections when learning a new language, but it is not always easy to spot. Suffice to say, I thought it was interesting enough to write this post.

Now, why did "-que" move to the end of words? Why did it stop being an independent word? Oh, the mysteries of Latin.

[Update: Lest you think I'm not sufficiently capable to talk about this in Latin...]

μέγας

I was poking about Wiktionary and found the page for μέγας. Don't ask. You don't want to know. Something struck me off the etymology section.

much (via OE michel) = magnus = μέγας

And I had never known that before. Isn't that cool?

Indirect Speech

Yeah, I think about grammar crud like this. Mostly because it came up in class last week.

And I had a realization: for once, English isn't the odd man out of the IE tribe.

In most Indo-European languages, indirect speech is pretty easy. Here's an example in English:

Direct: He's a good driver.
Indirect: I think he's a good driver.

The grammar machinery is pretty easy. Drop your head verb (which doubtless has a technical name) and your indirect-speech-itizer on your sentence, and presto! You're in indirect speech. Other IE languages do the same. Ancient Greek uses ὁτι, Persian uses که, English uses that and Spanish uses que. So easy. So simple.

But classical Latin can't do that. It has to do something very different. It has to press an accusative noun into duty as the subject of the indirect speech. The verb can't be a standard conjugated verb: it has to be an infinitive. So Latin, using our example above, does something like this:

Direct: Ille est gubernator bonus.
Direct: He's a good driver.
Indirect: Puto illum esse gubernatorem bonum.
Indirect: I think him to be a good driver.

And it's just weird. So congratulations to Latin for being the odd man out of the IE tribe for a change.

Note: As you see in the translations, English can do this bit of weirdness if needed. Ancient Greek can too. What's odd is that Latin doesn't have the easy "add in that for indirect speech" formula. Of course now I suppose I'll hear from all sorts of people telling me, "Dude, Armenian is just like Latin that way" or the like.