I’ve gotten a couple of review copies that I won’t have time to read for a while but that look so interesting and useful that I want to put the word out about them right away:
1) The Indo-European Controversy: Facts and Fallacies in Historical Linguistics, by Asya Pereltsvaig and Martin W. Lewis. The blurb says:
Over the past decade, a group of prolific and innovative evolutionary biologists has sought to reinvent historical linguistics through the use of phylogenetic and phylogeographical analysis, treating cognates like genes and conceptualizing the spread of languages in terms of the diffusion of viruses. Using these techniques, researchers claim to have located the origin of the Indo-European language family in Neolithic Anatolia, challenging the near-consensus view that it emerged in the grasslands north of the Black Sea thousands of years later. But despite its widespread celebration in the global media, this new approach fails to withstand scrutiny. As languages do not evolve like biological species and do not spread like viruses, the model produces incoherent results, contradicted by the empirical record at every turn. This book asserts that the origin and spread of languages must be examined primarily through the time-tested techniques of linguistic analysis, rather than those of evolutionary biology.
I’ve often cited Asya’s blog Languages Of The World (e.g., here), and that description is certainly singing my song (“the time-tested techniques of linguistic analysis”: hear, hear!), so I’m very much looking forward to reading the book.
2) May I Quote You on That?: A Guide to Grammar and Usage, by Stephen Spector. I generally look askance at popular usage guides, but this one looks very well done; it starts off pointing out that “We all use language in different styles, depending on the situation,” it’s written in a lively manner, it uses well-chosen quotes to make its points (and leads each section with them, rather than with a rule), and above all, it actually mentions the history behind the rules rather than presenting them as graven in stone — under “It’s or its,” for example, Spector says “The distinction that we use today didn’t become standard until the nineteenth century and into the twentieth.” For twelve bucks (or seven for the Kindle edition), you could do a lot worse.
The Indo-European Controversy does look interesting, but $100? *sigh*
Give it a few years and the price will drop to normal.
Actually, the price is “normal”, for a hardbound linguistics book, but as John says in a while a paperback will come out and it will be cheaper. In the meantime, there’s also the e-book for those who are into that sort of thing… Also, recommending it to your library might be a good idea…
Give it a few years and the price will drop to normal.
Years… or clicks on certain Russian websites.
Christopher Culver did an Amazon review. An excerpt:
“From even a cursory glance one could see that their model [Russell D. Gray & Quentin D. Atkinson] was extremely flawed, e.g. categorizing Polish as an East Slavic language, or dating the split of Romani from other Northern Indian languages to 1500 BC when it is known to have left the Subcontinent only around 1000 AD. Most linguists only shook their heads at yet another example of Nature and Science performing little due diligence with linguistics submissions, instead happy to publish sensationalistic findings. Unfortunately, the popular press took the 2012 paper up too, never mentioning that the scholarly consensus was overwhelmingly against it.”
As for the price, I’ve seen this a lot lately. Looks like when your book is pretty much destined for university libraries supply-and-demand does not apply.
Via MetaFilter: The trials and tribulations of translating Seinfeld episodes into German:
http://www.theverge.com/2015/6/24/8809723/jerry-seinfeld-tv-show-international-translation
The ebook is still over $80, and even used copies are over $70.
Culver’s remark about “links to blog posts at informal or amateur websites” being a weakness is less than gracious to either the authors or the bloggers in question, whoever they may be.
Also, from the book description: “the origin and spread of languages must be examined primarily through the time-tested techniques of linguistic analysis, rather than those of evolutionary biology.” The trouble with Gray & Atkinson is that they misapplied the techniques of evolutionary biology (as noted by our very own biologist) to data that was demonstrably incorrect. Abusus non tollit usum.
Give it a few years and the price will drop to normal
…. but there will be an entirely different set of geneticist’s conjunctions to critique … perhaps harder to critique because the gist of the geneticists interest in the ancient people-migration shifts away from the analysis of the phenomena of culture to the ancient DNA itself – but also because the paleogenetics quickly develops toolkit for analyses of admixtures / borrowings which will almost certainly engender new, better armed intrusions of the geneticists into historical linguistics, probably requiring an arms race of rebuttals.
John, in my review I wasn’t complaining so much about mentions of blogs by people working in the field (like, say, Languagelog, Amritas, Panchronica, or the authors’ own Geocurrents). The websites that I was disappointed to see presented as citations were those at the level of an old-school Geocities page put up by a possibly misinformed dilettante. It baffles me why Peretsvaig and Lewis thought it necessary to link to them when, with slightly little more effort, they could have linked to something rigorous. I can only imagine that the deadline was looming for the authors and, then, the editor wasn’t very involved.
Ah, okay. I withdraw my remark.
I wonder if I should try rewatching Seinfeld now that my English is somewhat better. It seems it went completely over my head back in the day (subbed, not dubbed as is the tradition here for anything other than TV for kids). Or perhaps my sense of humour is just too simple.
On the other hand, the Dolores/clitoris pun doesn’t work for me at all in in writing.
@Sili: I don’t think that joke works, period—because they changed the punch line. Apparently, Mulva’s name was originally supposed to be Chloris, which is funnier but also extremely uncommon. Somebody (I think from the studio audience) suggested Dolores instead, and they went with it.
For me it failed utterly, because Dolores has penultimate stress and clitoris has initial stress.
Same here.
Actually, the price is “normal”, for a hardbound linguistics book,
Yeah, that’s exactly the problem! I also think the book sounds great and I’m very happy that you got it into press, but it’s saddening to see dubious research promoted worldwide as a stunning breakthrough while counterarguments by people who actually work in linguistics are restricted to the plutocrat-and-university-library market because publishers can’t be bothered dealing with even the most conveniently assembled communities of interested amateurs.
(deep breath)
“Dolores has penultimate stress and clitoris has initial stress”
Yet, almost everyone I knew growing up pronounced them in a way that rhymed perfectly. The initial stress version still sounds clinical to me.
I learned “clitoris” with initial stress, but my wife gives it penultimate stress, and who am I to argue with her?
And by “failed utterly”, I meant not only that I didn’t find the joke funny, but that I found it utterly unintelligible and unguessable.
I have no idea where the claim about viruses comes from. The methods Gray & Atkinson were trying to use – phylogenetic analysis by Bayesian inference, and molecular dating – are routinely applied to all DNA and weren’t developed with viruses or any other particular group in mind.
One nit to pick:
Translated more literally, the joke would come across just fine. It would just be way too long for lip-synching: Wahnsinn – man sieht gar nicht, wo die Titanic damit zusammengestoßen ist… As the article goes on to say:
Surely German can’t be devoid of words for ‘hit, strike’ with less than six syllables.
Surely German can’t be devoid of words for ‘hit, strike’ with less than six syllables.
Well, if hit is supposed to mean “collide”, like in this case, the translations for “where the Titanic hit it” (7 syllables) I can come up with are:
1) wo er mit der Titanic zusammengestoßen ist – what David used (14 syllables)
2) wo er gegen die Titanic geprallt ist – (10 syllables)
3) wo ihn die Titanic gerammt hat – (9 syllables)
So, even the shortest possible translation has two syllables more than the English Version, cause you have to use the perfect in German here, which means auxiliary + verb in the past participle, and the past participle is at a minimum disyllabic.
Nouns are your friend. Surely “the scratch from Titanic” can be done in less syllables in German?
Would it be fair to say that English has an unusual economy of syllables among European languages?
Nouns are your friend. Surely “the scratch from Titanic” can be done in less syllables in German?
Der Kratzer von der Titanic. – 8 syllables.
wo ihn die Titanic gerammt hat
Is it possible to omit ihn here? The object is plainly recoverable from context, but I of course don’t know if German will tolerate it syntactically.
unusual economy of syllables
Supposedly written English is the tersest of the familiar European languages except for Hungarian, but I don’t know if this is really true. Translations in general tend to grow longer than their originals, because of the need to disambiguate what has been left vague in the original. In addition, English can pack a lot of phonemes into its syllables, with strengths (7 phonemes) the extreme case.
Supposedly written English is the tersest of the familiar European languages except for Hungarian
Huh. I checked that with the United Nations Universal Declaration of Human Rights, and sure enough, English used 1,772 words and Hungarian 1,567.
wo ihn die Titanic gerammt hat
Is it possible to omit ihn here? The object is plainly recoverable from context, but I of course don’t know if German will tolerate it syntactically.
Nope, not possible.
On the character level, however, the English UHDR has just over 11,000 characters, whereas the Hungarian version has over 13,000. (I think you omitted one word from your count of the English UHDR, most likely the word “Preamble”.)
That’s the one!
Huh. I checked that with the United Nations Universal Declaration of Human Rights, and sure enough, English used 1,772 words and Hungarian 1,567.
On the character level, however, the English UHDR has just over 11,000 characters, whereas the Hungarian version has over 13,000.
Bokmål: 1 637 (10 200)
Danish: 1 676 (10 800)
Swedish: 1 684 (10 900)
Nynorsk: 1 729 (9 900)
Faroese: 1 620 (9 800)
I can’t find the Icelandic version on Wikipedia.
Also:
Finnish: 1 307 (11 100)
Estonian: 1 419 (10 800)
(Rounded number of characters, exact number of words, as counted by Word.)
The Finns win!
Obligatory article here, object there, and unreasoned aversion to the good old plain preterite. Of course Danish dispensed with those Germanistic straitjackets long ago. Man kan ikke engang se hvor Titanic ramte = you can’t even see where the Titanic hit it at 12 syllables. (ikke engang is a three syllable word).
Of course agglutinative languages are going to have an unfair advantage here. I wonder what happens if you count morphemes, or phonemes, instead of words.
Characters are a decent approximation to phonemes for Hungarian, and an upper bound on phonemes for English, modulo the letter ‘x’.
The same can be said for Finnish versus the Scandinavian languages.
Nynorsk surprised me. I expected a short text, and so it is, but not that many more words than the other Scandinavian languages.
Counting phonemes in Article 1 of the UDHR I get 116 for English, 137 for Hungarian, so pretty much the opposite of the character ratio.
For these UDHR comparisons I think the character count you’d want is the shorter one excluding spaces between words, because what you are interested in correcting for is cross-linguistic variance in character-per-word ratio and you’re partially double-counting the words-as-such if you count the adjoining blank spaces. English on that measure (blindly trusting the methodology of the character count feature on my work word processor) is 8919 characters in 1773 words (using an approach yielding the same wordcount argued for upthread even though I might personally exclude the “preamble” and “article n” headings). Maltese (to take a European language that’s neither IE nor Uralic) gets down to 1528 words but uses 9714 characters. Welsh is marginally terser than English on both metrics (1727/8464), although both Irish and Scottish Gaelic are less terse than English on both metrics. Swahili (ranging farther afield) gets it down to 8294 characters (in 1729 words). The lowest character count I got (and I only randomly browsed a few languages I’m not mentioning here before getting bored) was Solomons Pidjin/Pijin, at 7797 (in 1809 words). Icelandic, which someone upthread was looking for without success, is only 1647 words, but 8635 characters (i.e. less than English but more than Welsh). I was using http://www.ohchr.org/EN/UDHR/Pages/SearchByLang.aspx rather than wikipedia as my source for texts, fwiw.
That was my source as well.
Die Finnen gewinnen!
(Instant rhyme illustrating why German usually takes more space to write than English, and longer to say.)
Less commonly used in such formal contexts, so it has to resort to periphrases?
No, not really. Well wriitten it’s terse and precise, as evident from the number of characters. The aversion to (or decomposition of) prefigated verbs increases the word count without increasing the number of characters much, but I didn’t expect the effect to be that big.
Interesting.
In Scots (and English varieties affected by the Northern Subject Rule) it would be “The Finns wins!”, which is obviously superior to all other alternatives.
Indisputable.
It may be of interest that zusammengestoßen has just three syllables in my dialect: zusammen- is one, -ge- is zero. But the rest of the phrase would stay uncompressed, giving a total of eleven.
zusammen- is one,
What would it be?
[t͡sɒ̈m]. In other words, the u and the e (both unstressed) are lost, and the resulting consonant clusters are maximally contracted. A very common and productive separable verb prefix; unlike in Standard German, it’s used only for the direction (“come together”, “eat up”…), not for the place (“be together”, “stay together”…), where it’s replaced by entirely uncontracted mitsammen or beisammen.
[t͡sɒ̈m]
Thanks, I would have guessed something like <zam>- (orthographically), but this is more precise.
It appears with actual [a] in Viennese mesolect, of course.
A very similar process leads to Schwyzertüütsch zäme, which however keeps the ⟨e⟩ as [ə], making me think that n-elision could have preceded apocope in Austro-Bavarian as well.
It doesn’t otherwise; it’s still there in infinitives and plurals and whatnots, unless another nasal precedes. Generally, Alemannic keeps the -e- and Bavarian keeps the -n of -en.
Edit: oh wait, loss of -n happens in other environments, too, in most dialects (not mine). The trick is that it never happens in the 1/3pl no matter if it happens in the infinitive; the 3pl still ended in -ent in Middle High German, so maybe -n was lost first, then the -t was lost, and then various levelings occurred.