I expected that my series on dogs would inspire a torrent of angry comments. After all, dog is one of the most enigmatic words in English etymology, but the responses were very few. I am, naturally, grateful to those who found it possible to say something about the subject I was discussing for five weeks, especially to those who liked the essays. As I have observed in the past, though I am supposed to love my enemies, I have a warmer feeling for those who are fond of me. At the moment, my dogs sleep in relative isolation, and that is where I’ll leave them. But suddenly my rather trivial old post on strawberry was picked up by MSN News, Hacker News, and Reddit, and thousands of people participated in the “chat.” Who could predict those five (three) minutes of evanescent fame? However, I also received a serious private letter. Our correspondent expressed surprise that I constantly refer to onomatopoeia and sound symbolism and asked me to clarify my attitude toward this matter. Summer is a dead season: outside the strawberry patch, in May and June almost nothing happened to me as an etymologist (that is why I even skipped my traditional gleanings last month—for the first time ever in more than ten years), and here was suddenly a consolation prize in the form of a big question. I am happy to answer it.
Those who are interested in word origins know the basic facts.
- For centuries people compared look-alikes and found worthwhile etymologies more or less by chance.
- In the nineteenth century, linguists discovered sound correspondences and learned to compare such words across languages that today bear little or even no outward resemblance to one another. The story began with Rasmus Rask (a Dane) and Jacob Grimm (a German) and continued with a group of mainly German scholars called Junggrammatiker in German and Neogrammarians in English.
- Quite often, words that violate regular sound correspondences still seem to be related, and no one knows what to do with them.
- Equally often the origin of words remains undiscovered despite numerous attempts to reconstruct their past.
First, it should be said that Jacob Grimm did not prove anything. Let us look at the most elementary example. We state that Latin p corresponds to Germanic f (this is “a law”), as evidenced by the pair pater ~ father. Then we formulate the rule that in our search of cognates we should be guided by the notion that if, for example, both Latin and Germanic words begin with f, they cannot be related (e.g. Old Engl. fæmne “woman”—long æ—and Latin fēmina). Either fæmne is a borrowing of the Latin noun (which is for many reasons unlikely) or it has an etymology of its own (which has been attacked more than once but never found). The circularity of the pater ~ father argument is obvious, and yet it seems that the conclusion is correct. To be sure, if our starting point were fēmina ~ fæmne (that is, always compare the Latin and Germanic words beginning with f and call them allied), the origin of fæmne would have been crystal clear, while the etymology of father would have presented an insoluble riddle. And yet, despite the fatal flaw behind Grimm’s reasoning, it appears that the results drawn from his premise are correct. Granted, they defy the main scientific principle, but they are worth salvaging!
The next example is of a similar type, but the reasoning is more convincing. According to the same law that pairs non-Germanic p and Germanic f, non-Germanic t corresponds to Germanic þ, that is, th as in Modern Engl. thin. But the Gothic for “father” (the oldest Germanic form available to us), was fadar, in which d sounded as ð (that is, as th in Modern Engl. this). Though the difference is minimal, Grimm and the Neogrammarians taught us to deal with sounds with utmost caution. Enter Karl Verner, another great Dane, who noticed that in the Sanskrit word for “father” stress fell on the second syllable (pitár). He checked numerous words and concluded that þ and other similar consonants (that is, fricatives, or spirants, as they are also called) were voiced if they followed, rather than preceded, an unstressed syllable. Consequently, Verner said, at one time even Germanic did not always have stress on the root. This discovery revolutionized Indo-European studies. But alas, dozens of words violate Verner’s Law!
Whence those “exceptions” to seemingly “exceptionless” laws? The answer is easy. Language is not elementary algebra, and hundreds of early and late words have individual histories. The rule of thumb is: try to apply the Neogrammarian principles to etymology. If they do not work, look for other factors, and you will discover sound symbolic and sound imitative formations, baby words, migratory words, taboo, and so forth. Your results will not be a hundred percent convincing, but that does not mean that they are wrong.
In this context, I want to tell a story that probably few people know. Half a century ago Jacques Rosenman, M.D. brought out a two-volume book titled Primitive Speech and English, which was followed by Onomatopoeia and Word Origins (1982). All three were published and distributed by the author. Rosenman noticed the circularity of Grimm’s Law and concluded that everything said by him and the Neogrammarians was nonsense; he treated Verner with special disdain. Rosenman used the worst tactic imaginable, for, ignorant of tons of special literature, he decided that no one had noticed the pitfalls of the “classical” theory. He screened numerous dictionaries and grammars (Noah Webster did the same at one time) but missed the fierce opponents of the Neogrammarians (such as Hugo Schuchardt), the researchers who insisted on the role of non-traditional factors in word formation (for example, Otto Jespersen, to mention the most famous name), and the huge body of literature on expressive sounds. Rosenbaum also missed Hensleigh Wedgwood’s English etymological dictionary. If he had studied that dictionary and reviews of it, he would have become aware of the many dangers his own approach entailed. Time and again he compared Indo-European and Semitic words, as though the groundbreaking books and articles by Hermann Möller, Alfredo Trombetti, Albert Cuny, and Graziado I. Ascoli had not existed. His books are full of reinvented wheels.
Rosenman also chose a self-defeating approach to his potential critics. Instead of showing that he was developing a fruitful view of etymology, he presented himself as an iconoclast and told the historical linguists whom he addressed in person that all past work had been a stupid mistake. As could be expected, no one took him seriously, especially because he was an outsider, and no journal agreed to review his work. Apparently, he never made it even to the section “Books Received.” He encountered only snobbery and at best puzzled indifference. I am telling this sad story not to add one more insulting remark to a host of those Rosenman had to endure. On the contrary, I believe that many etymologies he offered are correct and suggest that specialists make use of the three volumes he wrote: his material is rich, and his conclusions are often instructive.
Now to return to the last paragraph of the previous post. I said that every time I deal with words like big, pig, bug, bed, bad, dig, dog, and god (monosyllables beginning with and ending in stops), I end up with sound-imitative or symbolic formations. They indeed sometimes rebel against the Neogrammarian laws, but before one classifies them (or any other “recalcitrant” words) with “freaks,” it is necessary to exhaust all the traditional means of revealing their past. Jacob Grimm and Karl Verner have not gone to the dogs, but they were certainly not gods.
Image credits: (1) Lady Godiva by Adam van Noort, Public Domain via Wikimedia Commons (2) Cherchell museum – mosaic the bathers by Yves Jalabert, CC BY-SA 2.0 via Wikimedia Commons. (3) Karl Verner by Fr. Anders-Paltzow, Public Domain via Wikimedia Commons.
Featured image: Dog Sled Team by skeeze, Public Domain via Pixabay.
Scientists are not in the proof business: they leave that for mathematicians and creationists. Scientific laws are generalizations over facts, and often involve idealizations. Lead falls faster than feathers, but not because it is heavier.
I;m glad you’re still barking up the “dog” tree, because I have a comment regarding your analysis of Russ sobaka as a prefix so- + *baka. This felt strange, and sure enough my Russ edition of Fasmer analyzes the word as *sob- + an established pejorative suffix -aka. This leaves sob- as the problem, but I am trapped at home with nothing but Fasmer, so I can only think of o-soba, an individual. Could it have been used to refer to a big ol’ nasty dog? Why not?
I was wondering if hounds could be somehow related to sounds. Actually Latvian suns (old Genitive which became Nominative) lets think about other words with the same ending deguns (nose) which is rather strange formation (compare with nāss – nostril)… Actually short version of the radical for dog in Chinese reminds of the one for pig ( 犭- 豕) which both are very important animals for people. Latvian suns and sivēns (piglet) attest this similarity as well.