Undeception

The third of a three part series. See the others here:

Part 1: The Indo-Europeans

Part 2: The Tools of the Trade

____________________________________________________

Earlier I made mention of the consistency of sound changes, what the nineteenth-century German grammarians called the Ausnahmslosigkeit der Lautgesetze (the “exceptionlessness of sound change”) . The catalyst for this principle, which was in turn the catalyst for the existence of the discipline of historical linguistics, was the product of the work of two men, the first of whom was a German named Jakob Grimm (one of the Brothers Grimm who compiled the German folktales) in 1820.

Grimm, like others, recognized that the Germanic languages (among which are German, Dutch, Gothic, the Scandinavian languages, and English) share a common ancestor with Greek, Latin, and Sanskrit, based on a large number of vocabulary words whose phonetic similarities are too similar to be coincidental. Grimm’s important insight was that, in almost every word that appears to be related among Greek, Latin, Sanskrit, and the Germanic languages, words that had the consonants p, t, or k in the non-Germanic languages showed up in Germanic with the consonants f, þ (the “th” in think) or χ (as in German Bach) instead. He also noted that Gk/Lat/Skt b, d, g generally corresponded with Gmc p, t, k. The Gk, Lat, and Skt consonants that we now know come from Indo-European *bh, *dh, *gh he saw as corresponding with Gmc b, d, g (simplified here from a close phonetic variant). All three correspondences make up what we now call Grimm’s Law: in other words, he saw a method to the madness and was able to accurately predict the sorts of sounds that made Germanic the odd man out where Greek, Latin, and Sanskrit agreed with one another. Here’s an example.

Skt. pitar-, Lat. pater, and Gk. pater obviously correlate in some way with English father and German Vater (the “v” is pronounced like an f): so we have relationship between Gk/Lat/Skt p and the phonetically similar Gmc f, which was originally pronounced with both lips together. Likewise, Skt bhratar, whose first consonant bh continues Indo-European *bh, shows up in Germanic in the English word brother; bh vs. b.

Grimm’s Law explains one way in which Greek, Latin, and Sanskrit (most of all) better reflected the original Proto-Indo-European (PIE) pronunciation of the stops *p, *t, *k, *b, *d, *g, *bh, *dh, and *gh. One of the defining markers of the Germanic subgroup was this sound change that turned all those original consonants above into f, þ, χ, p, t, k, b, d, and g respectively. In tabular format:

First series: PIE *p, *t, *k > Gmc *f, *þ, *χ

Second series: PIE *b, d, *g > Gmc *p, *t, *k

Third series: PIE *bh, *dh, *gh > Gmc *b, *d, *g

This sound law accounts for many, many consonant differences between Gmc and Indo-European, but not all. The voiceless stops *p, *t, and *k occasionally did not show up as *f, *þ, and *χ as expected from the first series, but as *b, *d, and *g (the output of the third series)! It was not until 1875 that the Dane Karl Verner explained that all those exceptions to Grimm’s Law had their very own systematic explanation. The times when the PIE consonants did not undergo the expected sound change were cases in which 1) the consonant was not at the beginning of the word or 2) the vowel in the syllable immediately preceding the consonant carried the accent. This is now called Verner’s Law.

With the discovery that even the “exception” to the otherwise very regular (consistent) Grimm’s Law was itself regular, these early linguists were handed exciting evidence that sound change is carried out very regularly and systematically throughout the language, and hence can be reverse-engineered, “reconstructed” with a significant level of certainty that what is reconstructed will be accurate. Thus was the science of the comparative method born, and so the discipline of historical linguistics.

Is sound change truly exceptionless? Perhaps not, but no one ever really claimed it was: there are many cases in which the sound change for a few isolated words seems to have been reworked or replaced by other forms to make things seem uniform (the process called “analogy”). These exceptions are usually explainable on their own terms, and do not invalidate the sound change that took place on a broad scale for the vast majority of words; for instance, a particular word can be adopted from another language or dialect in which the sound change was absent or not as thorough-going. Nonetheless, instead of “exceptionlessness”, we now speak about the regularity of sound change that allows us to create sound laws that occasionally admit an exception here or there.

* In case anyone came down here to look for what the asterisks stood for, let me explain that unattested (reconstructed) sounds and words are marked by a preceding asterisk. Any word, root, or affix cited as being a PIE form should have an asterisk, since no one ever wrote anything down in PIE.

____________________________________________________

The third of a three part series. See the others here:

Part 1: The Indo-Europeans

Part 2: The Tools of the Trade

The second of a three part series. See the others here:

Part 1: The Indo-Europeans

Part 3: The Birth of Historical Linguistics

____________________________________________________

My last post dealt with the more anthropological side of my discipline. Most of what we know about the history of the Indo-European people groups comes not from historical records per se, but from analysis and comparison of the languages in which those historical records were composed. Philology (“love of words”) is an old term used to describe those who read literature for appreciation of the language. Naturally, most philologists of the eighteenth and nineteenth centuries were multilingual and well-read. It was this familiarity with multiple languages that led some to recognize similarities between the languages they were reading in, and as curiosity was given room, many of these philologists set out to figure out exactly the relationships between them. In so doing, they became intimately familiar with the language itself, not as it exists in practice, but in shared tendencies and similarities that must be chalked up to the innate nature of human language. Thus began the modern science of linguistics.

Because of ground-breaking work done in the 19th century, scholars discovered that sound change in language is a remarkably thoroughgoing and exceedingly systematic process. These linguists realized that if they could determine what sound changes occurred in a given language, they would be able to reverse engineer the process and uncover an earlier state of the language. That principle of reconstruction is what propels historical linguistics.

Systematic sound changes and changes in other aspects of language (such as in syntax, semantics, or morphology) are referred to as “laws”. For instance, a sound law that happened fairly recently in the transition from Middle to Modern English (c. 15th century) is known as the Great Vowel Shift, which was a drastic change in all the long vowels of Middle English and was responsible for the major differences between the way we pronounce and write our vowels and the way Spanish or other European languages do.

What historical linguists do is reconstruct language based on known laws using a combination of the comparative method and internal reconstruction.

*Disclaimer* I regret that not everyone reading this can read IPA phonetic notation. It’s a much more accurate way of representing the sounds than what I’m going to do. I hope those of you who know IPA will forgive me. It’s painful, I know.

The Comparative Method
The first part of the comparative method is determining whether two or more languages are genetically related (they share a common parent language). These languages may be singled out for comparison either because they share sufficient similarity (in vocabulary) or are supposed to be related through historical record. In the first case, one of the most important factors is the analysis of the similar vocabulary: are their vocabularies similar not because they come from a common mother language but because one language picked up some vocabulary through later contact with the other culture? Certain important types of words are not as subject to replacement by foreign words (usually the function words or others in the core of the lexicon like water or sun). Working from a reasonable certainty that the languages are genetically related, we look for differences in the similarities – yes, that’s what I said. In other words, we’ll take a word that is shared by both languages but is dissimilar in phonetic form and try to determine what they could both be derived from on the basis of known types of sound change elsewhere in human language.

Here’s an example. The first person to propose the counter-intuitive notion that the ancient Indian language Sanskrit was related to Greek and Latin was Sir William Jones, a British lawyer living in India in the late 18th century. He noticed that an astounding number of Sanskrit words were strikingly similar to Greek and Latin. This was the case with the word for “father”. Greek pater : Latin pater : Sanskrit pitar-. Many years later, by comparing this word with other shared words that have these same sounds in them, historical linguists postulated sound laws that explained the regular (systematic and predictable) system of how these “sister” or cognate words came to differ even though they had originally been identical in their shared mother tongue (Proto-Indo-European).

Internal Reconstruction
Whereas the comparative method looks between multiple languages, internal reconstruction looks within one language for relics of an earlier stage. For instance, consider the differences in wife and wives or bath and bathes. Notice that we pronounce the second consonant differently in wife and wives (f is voiceless and v is voiced), and even spell them differently. Although we spell both sounds “th” (don’t worry about why), bath and bathe similarly have a contrast between a voiceless and a voiced second consonant (“think” vs. “this”). Also keep in mind that <th> is not pronounced in either form as [t] plus [h], but as a single consonant.

Now, because each pair was derived from a single root, {wif} and {bath}, originally there must have been only one version of that consonant in both words of the pair, but something changed. The fact that we now have two versions of the consonant, voiced and voiceless, is a smoking gun that a sound change has occurred. Realize that the “e” in the spelling of both wives and bathe were originally pronounced as “eh” (it is just a spelling convention in the case of wife to tell us that the vowel is long). There was actually a stage of the language in which they were pronounced “wee-vehs” and “bay-theh“. If we set up the parallel, linguists notice something very telling. Let’s spell “wife” closer to the way it was pronounced, “wif”. So we have wif : wives and bath : bathe. What about the  surrounding sounds (“phonetic environment”) is a common denominator?

I’m glad you asked! In both cases, the voiced version of the second consonant sits between two vowels: -ife- and -athe-. What happened was that originally wives was pronounced “wee-fehs” and bathe was pronounced with a voiceless “th” (as in “think”) as “bah-theh”. At some point, the following sound change occurred: voiceless consonants became voiced between vowels. Then the “e” in the final syllables that didn’t have primary or secondary accent (emphasis) stopped being pronounced and dropped away except in spelling. But it was too late: the consonants were already being pronounced as voiceless in one form of the word and as voiced in the form that used to have a vowel after it. So this change in the phonetic environment masked the original conditions for the sound change (between vowels), except in how we spell it. Then at some point we started spelling “wifes” as “wives” to match the new pronunciation. If you made it through this far, congratulations: please post a comment with at least the word “woot” so I will know that you’re a trooper and that I haven’t completely wasted my time.

These are very confirmable examples, because Anglo-Saxon (the earliest written stage of English) actually spelled things closer to how they were pronounced, explicitly writing wives as “wifes” (showing the f pronunciation), for instance.

If this sounds way too esoteric or abstract, it is, I guess. But in me it excites my love of antiquity and my love for words. Even though I’ll probably never actually do anything important within this field (which, incidentally, goes far beyond simple sound changes), my understanding of the field helps me out as I analyze modern English and other languages today.

So there’s a little overview of some of the basics. Many pardons if I left you more confused than when you began reading!

____________________________________________________

The second of a three part series. See the others here:

Part 1: The Indo-Europeans

Part 3: The Birth of Historical Linguistics

The first of a three part series. See the others here:

Part 2: The Tools of the Trade

Part 3: The Birth of Historical Linguistics

____________________________________________________

Germanic and Indo-European studies. What the heck is that? Well, let me start with a summary of the anthropological side of the discipline. And as throughout the entire series, please forgive me for the gross oversimplifications.

Once upon a time, probably in an area along the steppes of Russia on the north side of the Black Sea, lived a people called the Indo-Europeans. They spoke a language we refer to as Proto-Indo-European (PIE). Gradually over the period of 3000-2000 BC, tribes within that culture began to migrate to other areas, taking their particular dialects of late PIE with them. Eventually, this culture ended up splintered into several groups that spanned from Europe in the west to India and the Xinjiang region of northwest China. As time passed and isolation from other tribes increased, each of these dialects became their own distinct languages and evolved as languages always do so that by the time of recorded history there was little real mutual intelligibility between the non-contiguous tribal dialects; these dialects became the branches, or language families, of Indo-European.

If the Indo-Europeans were not the first to domesticate the horse, they were at least among the first to make use of them. One particular branch known as the Indo-Iranians invaded what is now Iran and India with their horse power, where they became a conquering minority: although they were able to impose their language upon those they conquered, in time their other ethnic characteristics were lost among those of the peoples they ruled. In Iran, their descendants created the Persian empire; the Indo-Iranian language of the Persians lives on as Farsi today. In India, the language and many of the religious features of the Indo-Europeans lived on via Sanskrit, the ancient holy tongue of India, from which a large number of the many languages of India are descended (including Hindi). The fair-skinned Indo-Iranians called themselves “Aryans”, and it was the misconception of many early historical linguists that this name applied to all Indo-Europeans that led to the Nazi adoption of this term as a way if identifying themselves as descendants of a race supposedly dominant wherever it went.

Another group speaking what most scholars call Proto-Italic, or sometimes Italo-Celtic based on many features that suggest a close relationship with the Celtic languages (such as Welsh and Gaelic).  These speakers eventually made their way to Italy. Their language group evolved into the now-dead Oscan, and Umbrian, and the very much alive Latin, which lives on in the form of the Romance languages.

Another group, called the Hellenic branch, developed for the most part into Proto-Greek speakers and were the ancestors of the Myceneans and the Greeks, and apparently the biblical Philistines as well.

The language of the Balto-Slavic group split up into Lithuanian, Estonian, Latvian, and all the Slavic languages (Russian, Ukrainian, Polish, Belerussian).

The famous Hittites were Indo-Europeans, and were actually among the first to leave the homeland.

One more branch: Germanic. The subdivisions of the Germanic branch evolved into the extinct Gothic language, the Scandinavian languages (minus Finnish – it’s not really Scandinavian and is not Indo-European), German, English, and Dutch. This is my main field of study.

Other branches include Albanian and Armenian. All these branches should be thought of in some sense as sisters whose mother was Proto-Indo-European. These sisters became mothers of their own groups and languages, so that Spanish for instance is a child of Latin, which is a child of Proto-Italic, which is a child of PIE. Each of those “proto-”somethings is the nearest common ancestor of at least two languages: a mother language as it existed just before it split up into more than one daughter language is referred to as a proto-language. Another way to think of a proto-language is that it is the last reconstructable stage of a language before history records it. Proto-Italic was a language spoken until it broke up into what became the Latin-Faliscan and Sabellic (Oscan-Umbrian) branches. Proto-Latin-Faliscan was a language that broke up into Latin and Faliscan. Vulgar Latin, the language of the common people as opposed to formalized Latin, developed into Proto-Romance, a late stage of Latin that splintered into the different Romance languages.

That’s a macro shot of the endpoint of historical linguistics. I’ll explain a little about what exactly we do and how we go about knowing all this stuff in the next installment.

____________________________________________________

The first of a three part series. See the others here:

Part 2: The Tools of the Trade

Part 3: The Birth of Historical Linguistics