Grammar protects us
« previous post | next post »
Searching for something else, I happened across this quotation about language, attributed to the German (and later American) philosopher Eugen Rosenstock-Huessy (1888-1973):
Grammar and logic free language from being at the mercy of the tone of voice. Grammar protects us against misunderstanding the sound of an uttered name; logic protects us against what we say having double meaning.
I stared at these remarks with some astonishment. Have you heard of the Bulwer-Lytton Fiction Contest managed by the Department of English at San José State University? The one where people try to construct an opening sentence for the worst conceivable novel? The quotation above is like the winner of a bad writing contest where the task is to construct an opening sentence for the worst conceivable book about language and meaning.
Language Log could devote a thousand posts to the project of underlining and elaborating the ways in which grammar does not protect us against misunderstanding the sound of an uttered name, and logic does not protect us against what we say having double meaning. Come to think of it, the thousand posts may already have been written: there are over 5,500 old posts searchable here and already over 1,200 new ones on the present server searchable here.
Rosenstock-Huessy left Germany in 1933 (he was ethnically Jewish, but a Christian since his teens), taught at Harvard for a while (he was a bit too religious for them, it turned out), and then went on to teach at Dartmouth College until retirement. I don't know which of his works the above utterly-out-of-left-field quotation comes from (perhaps some learned commenter below will be able to enlighten me), but I would expect it is probably either from his book Speech and Reality (1970) or from The Origin of Speech (1981). The latter appears to be a defense of formal (i.e., public) as opposed to informal (intimate and interpersonal) language. According to the publisher:
Rosenstock-Huessy points out that the modern mentality has consistently preferred the informal to the formal, the abstract to the ritualistic, and numerical impartiality to personal address, and hence has forfeited the sources of a "grammatically healthy" community.
It is not easy to figure out what kind of view is being summarized here, or what grammatical health might be. I think I know what is meant by "informal" as opposed to "formal". These terms are normally used with regard to style (or what is sometimes known as register). Don't you think it's kind of chilly out? is informal Standard English, and Do you not believe that it is somewhat inclement outside? is formal. What is meant by "abstract" as opposed to "ritualistic" is much less clear to me. They don't seem to be mutually opposed. Language can be abstract whether or not it is ritualized, and ritualistic regardless of its abstractness. And as for "numerical impartiality" in opposition to "personal address"… I admit it, I do not understand what could possibly be meant by this. Every individual word is clear to me, but the way they are put together, the result (for me, anyway) is nothing.
The publishers add, "Communication scholars and linguists concur: no other writer has approached the problem of formality in language with the fresh insights to be found in The Origin of Speech." This may well be true. I am well aware that I could obtain the book and spend a few days studying it in order to find out. I might do that if (i) I were determined to understand everything that has ever been written about language, (ii) I had unlimited time available, and (iii) the author was not already known to have claimed that grammar protects us against misunderstanding the sound of an uttered name and logic protects us against what we say having double meaning. As things stand, however, I am going to have to leave it to you to read the book and tell me about it. Sorry, Rosenstock-Huessy may have been a very good man, but I can't read everything.
However, I wish you good grammatical health.
Mark Liberman said,
March 22, 2009 @ 11:25 am
This book attributes the quotation to the source "VAU" (p. 25), and says of VAU (p. 78) that
So maybe W.H. Auden made it up, or attributed to R-H something he heard or read elsewhere.
P. 25 also provides us with R-H's opinion that "God, World and Man are vocative and not nominative forms".
Simon Spero said,
March 22, 2009 @ 11:31 am
Any ambiguity may be the result of logic-free language.
Daniel von Brighoff said,
March 22, 2009 @ 11:49 am
Any ambiguity may be the result of logic-free language.
We have another contestant!
Dan Lufkin said,
March 22, 2009 @ 11:57 am
It could have been worse. According to the Amazon site, the following aphorism appears in seven of R-H's books:
The tree of everlasting life can grow only through successive generations of men reaching their hands to each other in one spirit across the ages. And each generation has to act differently precisely in order to represent the same thing. Only so can each become a full partner in the process of Making Man; only so can life be as authentic in the last age as in the first.
Don't count on me to review the book. I plan to wait for the movie.
Victoria Martin said,
March 22, 2009 @ 12:08 pm
Since Mr R-H is German, I'd guess that the opposition between "numeric impartiality" and "personal address" is a reference to the English pattern of using just one second person pronoun as opposed to the German choice between second person singular for personal address and second person plural for greater formality. But it's no more than a guess, since R-H seem to strive for opacity.
bianca steele said,
March 22, 2009 @ 12:10 pm
Free indirect discourse (FID) can be defined as a mode of speech and thought representation which relies on syntactic, lexical and pragmatic features.
"The quotation above is like the winner of a bad writing contest where the task is to construct an opening sentence for the worst conceivable book about language and meaning."
bianca steele said,
March 22, 2009 @ 12:30 pm
I'm actually fascinated by the history and sociology of accusations like R-H's: how the idea of opposing modernity to something else developed, how various ideas got attached to one side or the other of that opposition, how people have actually used the opposition and why, in their rhetoric — but I don't know of too many books that have examined this directly, rather than in passing, while discussing something different but related. I have been wondering whether a different subject matter than what I've been reading in might have more on the topic. There's "theory," but my impression so far has been of moving ideas around on the page without really confronting them deeply.
Brian Campbell said,
March 22, 2009 @ 12:58 pm
From The Origin of Speech, as found in the Amazon preview:
You really have to read this to believe it. He seems to be trying to claim that formal speech preceded informal speech; that informal speech is simply a bastardization of formal speech that somehow was created full-formed.
If you keep reading, he seems to be claiming that pronouns (and first names) are signs of this "informal" or "childish" speech that he's talking about. Don't ask how he writes formally without pronouns; he seems to use a normal proportion of pronouns, which according to his metric, must mean that he's speaking informally and childishly, if I understand his, er, logic.
Thankfully, I ran out of Amazon preview at this point, so I won't be subjected to more of this by virtue of train-wreck syndrome.
Ray Girvan said,
March 22, 2009 @ 1:08 pm
I just dipped into Eugen Rosenstock-Huessy: Studies in His Life and Thought at Google Books. Rosenstock-Huessy's idea of what grammar does is pretty unusual, lying in a theory that speech, not thought, is the basis for a Theory-of-Everything about human life.
It comes across as deeply cultured, erudite bilge.
Mary Kuhner said,
March 22, 2009 @ 1:10 pm
"Numerical impartiality" on its own would sound like the opening scene of _The Prisoner_:
"Who are you?"
"The new Number Two."
"Who is Number One?"
"You are Number Six."
"I am not a number, I am a free man!"
But then it is on the wrong side of the antithesis, as far as I can tell; more usually the numerical and impartial is also seen as formal and abstract.
This reminds me of an on-line debate I once had in which we were butting heads and getting absolutely nowhere. It finally dawned on me that when the other party to the debate said "concrete" he meant "perfectly specified, like a mathematical equation" and I meant "tangible and specific, like a rose." When he said "abstract" he meant "vague, nebulous, requiring interpretation"–like a rose, considered as a symbol–and I meant "generalized, conceptual" like a mathematical equation. In other words, we were fairly close to having opposite meanings for two key terms of the debate. If I recall correctly, I gave up at that point; it didn't seem likely that we could communicate. I don't think I could communicate with Dr. Rosenstock-Huessy either.
Stephen Jones said,
March 22, 2009 @ 1:47 pm
There is one very important point which people seem to be glossing over which is the importance of redundancy.
In informal speech "You want a coffee" can act as a question if it ends on a up-point and as a statement if the intonation is different (don't ask me to explain what happens in valley girl speech). In noisy environments the tone may be masked so the more formal version of the question "Do you want a coffee?" does make it clear it's a question independent of whether we catch the intonation, or whether the speaker uses an intonation pattern different from our own.
I fail to see how logic protects us against double meanings but grammatical gender certainly does. The fact that English lacks it explains why puns are so much more common in English than in French or German.
Janice Huth Byer said,
March 22, 2009 @ 2:39 pm
Mr. Pullum perfectly expresses my sentiments.
I read half way into the first sentence, "Grammar and logic free language from being at the mercy of tone of voice…" when common sense prevented me from pursuing further violations.
Since when is prosody a cruel bully and not the loyal servant of meaningful speech?
As a simple librarian, inspired by Language Log to fancy myself an aspiring autodidact in linguistics, I'm eager to be informed but dare not risk my tender knowledge base by reading anymore of Prof. Eugen, whose views strike me as too post-modern for my contemporary brain.
Simon Cauchi said,
March 22, 2009 @ 2:56 pm
@Janice Ruth Byer: "As a simple librarian, inspired by Language Log to fancy myself an aspiring autodidact in linguistics, . . ."
Join the club. Now, would you like me to send you a reading list?
Janice Huth Byer said,
March 22, 2009 @ 3:01 pm
Oops. I meant to write "Prof. Pullum" and "Mr. Eugen" NOT the obverse.
I apologize to Prof. Pullum for my confounding, which I'm tempted to blame on Mr. R-H for befogging my little mind with his ideas, but it's my bad. :)
Ray Girvan said,
March 22, 2009 @ 3:06 pm
too post-modern
It is very much like Baudrillard, but with an old-school Christian edge: Baudrillard cross-bred with Teilhard de Chardin.
Janice Huth Byer said,
March 22, 2009 @ 3:30 pm
@Simon Cauchi: "Join the club. Now, would you like me to send you my reading list?"
Yes, please! Thank you kindly. There's so much out there, I'd value help in separating the cream from the milk.
Not that the milk or even the whey has been wasted on a newbie like me. It was an MSM article that struck me as junk linguistics that led me to Google and landed me on Language Log, for which I'm forever indebted to that bad idea, since forgotten.
David Eddyshaw said,
March 22, 2009 @ 3:33 pm
"Grammar and logic free language from being at the mercy of the tone of voice."
Rosenstock-Huessy might be thinking primarily of written language (he would appear to be the sort of person who would think that way).
It's true enough that spoken English transcribed into written form suffers from the stripping out of all the information about stress and intonation and pauses which cannot be represented by the blunt instrument of traditional punctuation, and that even informal written English is constrained by this in ways which Rosenstock-Huessy would probably interpret as being "more grammatical".
peter said,
March 22, 2009 @ 3:44 pm
"Grammar protects us against misunderstanding the sound of an uttered name; logic protects us against what we say having double meaning. "
E R-H seems to think that the allegedly-precluded sonic misunderstandings and double meanings are undesirable. How quaint!
Neal Goldfarb said,
March 22, 2009 @ 3:51 pm
@ Stephen Jones: "I fail to see how logic protects us against double meanings but grammatical gender certainly does. The fact that English lacks it explains why puns are so much more common in English than in French or German."
I have to admit that I'm more than a little confused as to (1) how grammatical gender protects against double meanings, (2) what grammatical gender has to do with the frequency of puns in a particular language, (3) how the frequency of puns in a particular language can be determined, and (4) the basis for saying that puns are more common in English than in French or German.
I'd also be curious to hear an English-language pun that turns on the absence of gender-marking, since I can't think of any offhand.
J. W. Brewer said,
March 22, 2009 @ 3:59 pm
The thing about "numerical impartiality" I take to probably imply a fairly common stock critique of modernity, i.e. that it leads people to deal with each other in a cold, rationalistic, and scientistic (thus falsely objective and "impartial") way, with the characteristic scientistic flaw of treating those aspects of experience that cannot readily be quantified and thus specified numerically as irrelevant. This is then contrasted to a more humane and and "relational" mode of interaction supposedly characteristic of the premodern period. I think Martin Buber might be a representative Germanic exponent of this critique from the generation one or two prior to R-H. At least he'd be one who used to be reasonably widely read in English translation.
I'm not sure the thing about speech in some sense being more important than thought is that weird. Isn't one of the consequences of the private language argument (at least when added to other bits of Wittgenstein) that we really can't conceive of what it would feel or be like to be a thinking/self-aware creature with no capacity for language?
I admire the audacity of the image of a Baudrillard/Chardin hybrid, but it makes my head hurt.
Daniel von Brighoff said,
March 22, 2009 @ 4:02 pm
I fail to see how logic protects us against double meanings but grammatical gender certainly does. The fact that English lacks it explains why puns are so much more common in English than in French or German.
Unless you've solid proof that puns are, in point of fact, more common in English than in French or German, you're very much begging the question here. I'm not even sure how one would go about collecting such data in an empirically sound manner, but it's got to be simpler than establishing any sort of causal connexion between grammatical gender and punning.
Nathan Myers said,
March 22, 2009 @ 4:13 pm
One suspects that "numerical impartiality" refers to the scrupulously observant use of numerical pronouns, such as "one", where a personal pronoun such as "you" or "I" might be unnecessarily overspecific. If such a suspicion were to bear out, one may reasonably conclude that one's hope to find evidence for the coherent erudition of one's putative better dashed once again.
Stephen Jones said,
March 22, 2009 @ 4:32 pm
I think a study of newspaper headlines in French, Spanish and English would provide a fair amount of proof that puns are more common in English.
With regard to ambiguity (which is a necessary condition for a pun) I would have thought it clear that one of the main purposes of grammatical gender is to provide redundancy. The fact that both the noun and the adjective would normally need the same ending would make the confusion about what part of speech a word is in English, and thus confusion over the overall meaning of the utterance, much less likely.
To check the theory out you could take a few hundred sentences in English, and their equivalents in Spanish and French. Then look at the phonetic transcripts and count the possible amibiguities. Then, which is my point about redundancy, make one change and the phonetic transcript and look for the possible ambiguities, and continue to do so for each phoneme.
Another way of looking at it is the theoretical way. In a language without gender distinction a noun can normally be confused by one phoneme change with a certain number of other words in the whole list of nouns. If there are two genders it can only be confused with the same proportion of words in half the list of nouns, which reduces the possibility of ambiguity by more than half.
Stephen Jones said,
March 22, 2009 @ 4:42 pm
To give you an example. Take the Catalan noun la paret, 'wall'. Now if you don't hear the 't' at the end then you hear the word pare which is 'father' but you will not misintepret what you've heard because fathers are almost never feminine.
Terry Collmann said,
March 22, 2009 @ 4:54 pm
I think a study of newspaper headlines in French, Spanish and English would provide a fair amount of proof that puns are more common in English.
No, it could only prove that puns are more common in English newspaper headlines.
I would have thought it clear that one of the main purposes of grammatical gender is to provide redundancy … there are two genders it can only be confused with the same proportion of words in half the list of nouns, which reduces the possibility of ambiguity by more than half.
If you're trying to say that the speakers of PIE sat around saying: "Our language has too many ambiguities: we'll have to introduce grammatical gender to cut down on the confusion," I'm afraid I'd have to demur …
Stephen Jones said,
March 22, 2009 @ 5:01 pm
There's also the question of where Spanish for example will have one word from two entirely separate derivations.
We have el fondo, 'the bottom', from Latin,and la fonda, 'the inn', from Arabic. The separate genders makes them clearly distinct, whilst I suspect in English we would have ended up with identically sounding words.
Stephen Jones said,
March 22, 2009 @ 5:12 pm
No, don't put words into my mouth. What I am saying is that if a feature in a language doesn't have an effect it tends to disappear. If you want to, you can believe, with Bryson and others, that gender only persists in languages as the result of some primitive sexist or animist classification of the world, but as nobody seriously suggests the present gender classification of French or Spanish or German represents the French, Spanish or German speakers classification of reality, then there must be some advantage to maintaining the classification.
True but we have to ask the question why, and I believe the reason is that it's so much easier to do in English.I suggest you take the days news and try and make up punning headlines in French, Spanish or German. I swear you'll find it a hell of a lot more difficult than doing it in English.
Stephen Jones said,
March 22, 2009 @ 5:15 pm
Or you could trying comparing the density of puns in Shakespeare to those in Lope de Vega or Calderón de la Barca. There are various aspects of English grammar (of which the lack of grammatical gender is only one) that makes word play easier (and thus ambiguity more likely) than in other languages.
Stephen Jones said,
March 22, 2009 @ 5:21 pm
To put it another way a pun necessarily requires an ambiguity. And if there are more ambiguities in English there are more possiblities of puns.
Some languages have a grammatical structure that lends itself to ambiguity more than others (in general the greater the overt grammatical complexity the less the opportunity for ambiguity).
Simon Cauchi said,
March 22, 2009 @ 5:33 pm
@Janice Huth Byer: "Yes, please! Thank you kindly. There's so much out there, I'd value help in separating the cream from the milk."
Try these for starters, undergraduate textbooks that can be read with pleasure:
English in Australia and New Zealand, by Kate Burridge and Jean Mulder (Oxford UP; don't be put off by the Antipodean title; the book can be read with profit anywhere)
An Introduction to Historical Linguistics, by Terry Crowley (Oxford UP; my copy is of the 3rd edn; for all I know there may be a later one now; the examples are mainly from Australian and Pacific languages)
The English Language: A Historical Introduction, by Charles Barber (Cambridge UP; it starts with a general overview, then narrows its focus to Indo-European languages, then Germanic, finally English)
and for light relief go on to
Language Myths, edited by Laurie Bauer and Peter Trudgill (Penguin)
Blooming English, by Kate Burridge (ABC Books; essays, or rather printed versions of broadcast talks; ABC stands for the Australian Broadcasting Corporation)
and for an ambitious overview of language in general
A History of Language, by Steven Roger Fischer (London, Reaktion Books)
Lance said,
March 22, 2009 @ 5:49 pm
Don't get distracted by the discussion on the origins and relative commonness of puns, folks. Instead, be sure to follow Ray's link above and read for yourself the first few pages from which he excerpts. He's not exaggerating when he says that Rosenstock-Huessy considers pronouns to be part of "informal" language (as well as nicknames like "Jim"; full names are formal language). "Formal", he makes clear, covers
Linguists, of course, are entirely wrong to study "informal" language, like that between child and mother or officer and platoon. Those are degraded, and—I can't even go on, it's too ludicrous for me to begin to believe. I'm with Ray—pardon me, with Mr. Girvan; it's a good thing the Amazon preview didn't go any farther, or I'd still be turning virtual pages in stunned disbelief.
Skullturf Q. Beavispants said,
March 22, 2009 @ 6:42 pm
Regarding the question of grammatical gender of nouns, and the practicality thereof, or the tendency thereof to stay in a language and not disappear:
Quine once pointed out one possible advantage of grammatical gender. The English sentence "Mullins removed the manuscript from the briefcase and threw it into the sea" is ambiguous. What was the "it" that Mullins threw into the sea: the manuscript, or the briefcase?
If "manuscript" and "briefcase" are two different genders using two different equivalents of "it", you no longer have that ambiguity.
acilius said,
March 22, 2009 @ 6:42 pm
I haven't read Rosenstock-Huessy, and most of the excerpts above don't sound promising. One sentence that I'd be willing to stick up for is this: "Nobody could say Mommy or Daddy unless someone had spoken reverently Father and Mother before." Well, that may actually be true, don't you think? After all, where is the point of such familiar diminutives as "Mommy" and "Daddy" if not in their contrast of register with "Mother" and "Father"?
In general, yes, we can say that formality is as much a negation of informality as vice versa, and so we are not warranted to say that formal speech is older than informal speech. The two categories must have come into being simultaneously. But, as far as "Mommy" and "Daddy" go, I think he may be onto something.
Perhaps the Jakobsonian argument that parents have simply seized on "mama" and "papa" forms as baby talk for "mother" and "father" can be connected to this. Remember the late Larry Trask's little story about "Jennifer's word" in his famous "Where do mama and papa words come from?" Parents, eager to hear their child begin to speak, interpret babbling as babyish words. When baby Jennifer babbles "ma" sounds, "mama" is declared to be "Jennifer's word for mother." It is just because the word is different from the formal word that this interpretation hassuch a hold over the mind of the parent.
Of course, informal words of this sort sometimes crowd out more formal words, as indeed "mom" seems to be well on its way to crowding "mother" out of use in American English. In today's USA, "mother" seems to be taking on an ever-more formal resonance as "mom" proceeds from strength to strength.
Morten Jonsson said,
March 22, 2009 @ 7:56 pm
Skullturf Q. Beavispants said: "If 'manuscript' and 'briefcase' are two different genders using two different equivalents of 'it', you no longer have that ambiguity."
Perhaps, but they might very well be the same gender, in which case the potential ambiguity might be even more marked. Removing two-thirds or three-quarters of such ambiguities (depending how many genders there are) might be convenient, but I don't know if it's really a compelling reason for retaining gender. Nor are the other nifty things you can do, such as (in German) using definite articles as personal pronouns. They're part of how a language works, but surely they're not in themselves the reason it works that way.
And as far as ambiguity, there are always strategies within a language for dealing with it; the result may be more awkward than what you wanted to say, but surprisingly often, with a little thought, it will actually turn out better–more graceful, more natural. I'm still working on a graceful version of that manuscript/briefcase one, though.
Nathan Myers said,
March 22, 2009 @ 8:14 pm
acilius demonstrates once again that there is no proposition so silly that it cannot find an earnest, or at least a game, defender. I wonder, though, whether acilius himself will report first having spoken the formal or the diminutive maternal noun in his milk tongue.
The question of punning in newspaper headlines might be further illuminated by considering the record of Mandarin and, particularly, Japanese newspapers. Supposedly puns are much easier in those languages, and much of Japanese literature depends heavily on punned ambiguities.
Aleksei Nazarov said,
March 22, 2009 @ 8:35 pm
Actually, the first part of the first quotation from R.-H. doesn't seem that bad to me (I mean the first sentence "Grammar and logic free language from being at the mercy of the tone of voice."), provided that it is set apart from the naive and nonsensical ideas that R.-H. seemed to have about language.
I interpret this quote as follows: "grammar and logic" contribute certain elements to language expressions that are independent from purely pragmatic and paralinguistic considerations ("tone of voice"), which allows us to linguistically express content that is contradictory to the "way we say it".
Thus, this quote simply seems to state the distinction between semantics and pragmatics.
marie-lucie said,
March 22, 2009 @ 10:31 pm
Mommy/Daddy versus mother/father:
Among other misconceptions, R-H seems woefully ignorant of basic facts of language development and history. The terms of address did not spring ready made from contemporary English (and the same is true for their equivalents in other languages), but have a long history. Baby talk is particularly resilient, resisting changes which occur in regular words, and in any case Mommy/Daddy are not formed directly from mother/father: no one says to their baby "say Mother" or "say Father", from which the baby comes up with Mommy or Daddy – those terms of address are what the baby hears from the parents. Also, the -ther component of the English words corresponds to -ter in Latin, and both are from a PIE suffix attached to roots similar to ma and pa, so that the ancestor words were complex, even though their present descendants are perceived as simple. The first root ma could have been influenced by the sounds a baby can produce, but in some languages mama means Daddy rather than Mommy, and in many other languages the regular words for parents do not have "easy sounds" pronounceable by an infant.
The briefcase and the manuscript:
The sentence as written is indeed ambiguous, but if it was part of a larger text such as a short story or novel there would probably be some earlier context which would indicate which of the two objects needed to be gotten rid of, and/or perhaps some later context indicating which one remained with the person who threw the other one into the sea.
Puns in various languages:
French and English puns are based on different principles, corresponding to their respective phonological and phonotactic structures.. Since there are lots of homophones in English, especially in one-syllable words, it is possible to play on the various senses of some of the written words. This is especially so in newspaper headlines which tend to be written in telegraphic style, omitting grammatical words and therefore making the relationship of the words to each other in the sentence ambiguous. In French, because of the syllable-based rhythm and the phonetic linkage of words to each other in phrases ("liaison"), something which tends to blur the limits of individual words even if those words are not inflected, much punning is based on the fact that the same phonetic sequence of syllables can often correspond to different morpheme and word sequences. A written version is unambiguous, so you don't often see such puns written (at least not in newspapers), since the ambiguity only arises from the pronunciation, and the visual appearance of individual words meant to be read silently does not necessarily call to mind the potential ambiguity. But you do sometimes see such puns in humorous works or satirical papers such as Le Canard Enchaîné, where readers expect them.
Catalan paret 'wall' and pare 'father': isn't there a difference in stress, on the second syllable in paret but the first syllable in pare, so that the words could not easily be confused in speech even under less than ideal conditions?
J. W. Brewer said,
March 22, 2009 @ 10:37 pm
Is there an extant language (or even a well-attested defunct one) anywhere without some distinction between formal & informal registers? The gap in English (for some subsets of speakers, including me), may be not as wide as in many other speech communities — you can't really call what we (or at least I) have diglossia and we don't have a whole separate set of grammatical rules deployed only when interacting with the imperial family. Heck, we don't even have a T/V distinction in second person pronouns. But we still convey formal v. informal in various syntactic, phonological, and lexical ways. Is there a language which doesn't? Even Piraha? (Pidgens maybe shouldn't be eligible, nor should languages whose living use is now confined to a liturgical or similarly specific context — I mean a language that is a first language for most of its speakers and whose speakers generally use it in the full range of social situations they may expect to encounter.)
Jonathan said,
March 22, 2009 @ 11:00 pm
Les lettres du blanc sur les bandes du vieux billard. /Les lettres du blanc sur les bandes du vieux pillard.
That's a wonderful pun from the writings of Raymond Roussel, where the two phrases differ by the voicing of a single phoneme, but whose meanings are rather distinct:
"Letters [alphabetic symbols] written in white on the sides of the old billiard table" versus "Letters [correspondence] written by a white man about the armies of the old pillager."
I don't think gender is a serious obstacle to the creation of puns, based on my experience of various romance languages. In French, especially, a lot of words different words are pronounced the same, like vers (worms), vers (towards), vers (lines of poetry)… which sound almost like "verre"
You could also probably make puns with words with different meanings in masculine and feminine, like in Spanish "el orden" vs. "la orden" or "la radio" vs. "el radio."
In fact, I could invent a very stupid joke:
"Por qué no se casó el Papa? –Porque no había papas."
marie-lucie said,
March 22, 2009 @ 11:20 pm
Les lettres du blanc does not mean "letters written in white", which would be les lettres en blanc. I find the interpretations of both sentences tirées par les cheveux. Surely there are better puns that this one.
rootlesscosmo said,
March 22, 2009 @ 11:38 pm
My high school French teacher quoted one about the driver of a donkey-cart:
"Je ne suis pas çe que je suis, car si je suis çe que je suis, je ne suis pas çe que je suis." I wouldn't call it a knee-slapper, but it's a pun, isn't it?
Skullturf Q. Beavispants said,
March 22, 2009 @ 11:49 pm
I admit that Quine probably wasn't asserting that the manuscript/briefcase example illustrates in a historical sense "why" grammatical gender has been preserved, and probably meant it more as a cute illustration of an unexpected feature of grammatical gender.
Re puns in French and the blurring of the boundaries between words: I remember being introduced to a joke when learning French to the effect of "how do you make a cow become worth a thousand francs?" The answer involved feeding the cow one franc, then giving her something to make her throw up, and then:
Elle vomit le franc, which sounds like Elle vaut mille francs.
marie-lucie said,
March 23, 2009 @ 12:39 am
My high school French teacher quoted one about the driver of a donkey-cart:
"Je ne suis pas çe que je suis, car si je suis çe que je suis, je ne suis pas çe que je suis."
The rhyme I know is spoken by a footman walking behind his master (the driver of a cart does not really follow the animal as only the animal walks or trots):
Je ne suis pas ce que je suis,
Car si j'étais ce que je suis,
Je ne serais pas ce que je suis.
The pun is caused by the homophony between je suis 'I am' (verb être) and je suis 'I follow' (verb suivre). It only works in the first person singular.
Elle vomit le franc, which sounds like Elle vaut mille francs.
The pun works only if you pronounce vo- and vaut the same, which I don't, any more than most speakers (originating) from Northern France. But it works in Southern French, in spite of a slight stress difference which differentiates the two sentences.
Stephen Jones said,
March 23, 2009 @ 2:29 am
Marie-Lucie. Correct about stress on pare and paret; the final vowel may be different depending on the variant of Catalan you speak. The point I am making though is that there is one more protection against misunderstanding. People fail to recognize that redundancy is a good thing in languages as well as in protection measures for nuclear reactors.
Karen said,
March 23, 2009 @ 6:24 am
At the risk of a bit of information restructuring:
Mullins threw the manuscript into the sea after removing it from the briefcase./ Mullins threw the briefcase into the sea after removing the manuscript from it.
Or fronting the 'removing' clause:
Before he threw the briefcase into the sea, Mullins removed the manuscript from it. / Before he threw the manuscript into the sea, Mullins removed it from the briefcase.
Or, changing the reference pronoun to a co-referent:
Mullins removed the manuscript from the briefcase and threw the papers into the sea. / Mullins removed the manuscript from the briefcase and threw the case into the sea.
As pointed out above, ambiguity is still possible with gendered pronouns. What's worse is that as a translation teacher. I'm always having to remind my students that if a feminine "it" in Russian has a neuter or masculine noun between it and its antecedent, their English translation will probably *not* be able to use "it", because it will in English not refer back to the same noun. You'd think they would notice that, but they often don't – probably because, as do the writers of original ambiguous sentences, they *know* what it means.
Peter Erwin said,
March 23, 2009 @ 7:40 am
At this point I'm willing to bet (though not a lot) that somewhere else on the Internet is a discussion where someone is earnestly asserting that, e.g., the smaller number of phonemes in Spanish compared to English means that ambiguity is easier in Spanish, and that therefore there are fewer puns in English…[*]
The presence of grammatical gender in Spanish doesn't seem to interfere with making puns in Mexico, if this post is any indication: "Puns are the staple of Mexican humor. People use them on a daily basis, and these puns, or albures, are almost always sexual jokes or disguised insults." (Follow the link for some examples.)
As other people have pointed out, the presence or absence of puns in newspaper headlines is apt to tell you more about the individual microcultures of headline writing than about the languages themselves — whether one is allowed (or supposed) to commit puns in headlines, rather than how easy or difficult it is. I suspect[**] that one would find a different frequency of headline puns if one stuck to "highbrow" newspapers like The NY Times, The Washington Post, etc., than if one looked at The NY Post and its ilk.
[*] Without providing any actual evidence for this, of course.
[**] But of course haven't done any research on this.
acilius said,
March 23, 2009 @ 8:14 am
@nathan myers: I suspect that when R-H says that formal language is older than informal language, he does not mean that individuals must master a formal register before they can use an informal one. Rather, I suspect he means that language itself originated as a solemn and esoteric ritual structure and that informal registers were introduced as a rebellion against this solemnity. Of course, this idea strikes most linguists today as very silly, since it seems far more likely that language is an evolved capacity than an invented technology. But if we can imagine a time when it seemed plausible that language might have been an invention of humans, I think we can see where R-H was coming from.
bianca steele said,
March 23, 2009 @ 8:31 am
Regarding Quine's example: It seems to me the natural reading of Quine's example is plainly that the manuscript was thrown into the sea, the focus having already been on "manuscript" in the first part of the sentence. A language with gender might permit the focus to shift to the briefcase without changing the word order (or the wording, as seen from the perspective of an English speaker). English could also use "the former" and "the latter," but these are very formal.
marie-lucie said,
March 23, 2009 @ 9:08 am
SJ: I would have thought it clear that one of the main purposes of grammatical gender is to provide redundancy.
In that case, one has to wonder about why English lost the category in nouns and did not make it up in other ways (addition of specific morphemes such as in he-goat or actress is extremely restricted, applies to "biological gender" not grammatical gender, and -ess seems to be on its way out).
The fact that both the noun and the adjective would normally need the same ending …
That is called agreement, same as with subject and verb, and it could apply to case and number alone even without the existence of grammatical gender, as in the English pronoun they/them. The ending does not have to be the same: Latin with its three major declensions (the other two being marginal) corresponding to noun-classes (one of which includes words from all three genders) often has nouns and adjectives matching in gender and case but not in specific endings, and there is frequently some overlap between different endings in different classes (eg -is for both genitive singular in the 3rd declension and dative/ablative plural in the 1st and 2nd declensions – length not being written in ordinary Latin), something which makes for many possible puns in Latin, as for instance in mottoes, which are very often ambiguous.
marie-lucie said,
March 23, 2009 @ 9:11 am
Karen: of course the sentence with the manuscript and the briefcase was not the clearest, but the point was not how to rewrite it (all your suggestions were very good) but how to arrive at an unambiguous interpretation by adding context.
Ray Girvan said,
March 23, 2009 @ 9:13 am
My favourite French pun is the one about "La Vie en rose" translating as "the pink aeroplane" (L'avion rose).
Skullturf Q. Beavispants said,
March 23, 2009 @ 9:40 am
"I'm very hungry" in two letters: G a
(G grand, A petit = J'ai grand appetit)
Chris said,
March 23, 2009 @ 9:49 am
But if we can imagine a time when it seemed plausible that language might have been an invention of humans, I think we can see where R-H was coming from.
It seems to me that this would compel the opposite conclusion. Human inventions nearly invariably start out crude and improve in successive generations; compare the Lascaux cave paintings to the Mona Lisa, stone axes to Solingen steel, dugouts or hide canoes to a galleon (to say nothing of an aircraft carrier!)
Thus, this logic would lead to the conclusion that language started out as little more than a collection of grunts without formalities, and the complexities of modern language were developed later. Techniques, no less than technologies, do not spring fully armed from the skull of Jupiter.
peter said,
March 23, 2009 @ 9:51 am
Stephen Jones said: " would have thought it clear that one of the main purposes of grammatical gender is to provide redundancy.
This is not at all clear to me. Why is or was redundancy desirable in a language? If this property is (or was) desirable, why do some languages have just two genders and others (eg, those in the Southern Bantu family) many more? What added benefits do the 21 or so noun classes of chiShona provide over French's two classes? Any explanatory theory of noun classes has a lot more to explain than merely the genders of French and Spanish.
Aaron Davies said,
March 23, 2009 @ 9:57 am
anybody know how common punning is in polysynthetic languages?
all the french "puns" i can think of at the moment are usually categorized as tongue-twisters instead–"Napoléon, cédant Sedan, céda ses dents", etc.
japanese is indeed easy to pun in, at least judging by anime dialogue, and is presumably helped both by a relatively analytic syntax and an extremely limited phonetic repertoire. several article have appeared here recently about tone puns in chinese used to get around censorship.
norman fiering said,
March 23, 2009 @ 10:58 am
Astonishing academic condescension on display here. All that superior commentary without anyone having read Rosenstock-Huessy's Origin of Speech, which is an unconventional book but truly a masterpiece. Rosenstock-Huessy was no crank. Check him out in the Stanford Encyclopedia of Philosophy. He was a person of immense learning and the deepest seriousness who wrote in many different fields. Although he spent his life mostly within academe, he was appalled at the blinders most "authorities" wore and at their timidity in the face of any challenge to the reigning paradigm. Circle the wagons. Someone is saying something you have not heard before.
[(myl) With respect, the issue here is not that R-H is saying something that we haven't heard before, but that he's saying something that appears to be complete nonsense. If we're wrong, please translate the cited quotes so that we can understand their value.
I don't think that reading Origin of Speech is going to help us. I've read the first chapter, and it strikes me as nonsense piled on nonsense — and not, as far as I can tell, because its ideas are new to me. On the contrary, the most nonsensical parts seem to me to be the most conventional ones. ]
acilius said,
March 23, 2009 @ 10:59 am
@Chris: "It seems to me that this would compel the opposite conclusion. Human inventions nearly invariably start out crude and improve in successive generations." If language had been invented, we would certainly have expected it to have been crude at first and to have become more complex over time. However, that wouldn't have kept language from starting as a solemn ritual structure. Rather the contrary, I should think.
If language had been a human invention, as in Rosenstock-Huessy's prime many people believed it to be, we would expect the first generations of language users to learn only a crude version of it, and that only through frequent repetition. Probably that first language would be so crude that it would have little or no obvious utility. Why repeat phrases over and again if they are of so little use? Surely the first answer that would come to mind would be that the phrases were part of a sacred ritual.
Likewise, if we believed that all grammatical rules were invented, then we would believe that for each grammatical rule, there was a time when that rule was known only to its inventor. What process would that individual have had to institute to teach a rule to the rest of his or her community? Surely it would have to be a process that was repeated frequently and in which many people were expected to take part. So, again, the first explanation of that process that would be likely to come to mind would be that it was a sacred ritual of some kind.
While few scholars today would defend the notion that language was a human invention, there's nothing old-fashioned about seeking the origin of other aspects of human behavior in ritual. So, to borrow one of your examples, the Lascaux paintings are often interpreted as cult artifacts, and the visual arts are commonly thought to have emerged in ritual contexts long before they were developed for non-ritual purposes.
Stephen Jones said,
March 23, 2009 @ 1:02 pm
I am not suggesting that noun classes arose to increase redundancy. I am merely pointing out that that is one of their side effects.
Redundancy is obviously a good thing in language. If there was no redundancy every time you heard something in less than perfect conditions your understanding would seriously suffer. The question is exactly how much redundancy is a good thing. It seems to me that there is something of a band here, with some languages having more redundancy than others.
Stephen Jones said,
March 23, 2009 @ 1:07 pm
I was always under the impression that English lost its genders because the Norman French speakers had one set of genders and the English speakers had another, which is why English also lost most of its declensions and conjugations.
Stephen Jones said,
March 23, 2009 @ 1:16 pm
It would of course depend on mishearing the phoneme that is different in the two cases.
And it is quite possible they would have a very good case (though the number of vowel phonemes in English varies wildly according to the geographical variant, and the distribution is far from equal, with the schwa being much more numerous than any other phoneme). That there is one factor that makes ambiguity more likely in Spanish than English doesn't preclude that there is a factor that has exactly the reverse effect.
Stephen Jones said,
March 23, 2009 @ 1:24 pm
Correct. And agreement strengthens redundancy. A language with genders will often have noun-adjective agreement and thus more protection against misinterpretation, than one that doesn't.
Nathan Myers said,
March 23, 2009 @ 2:19 pm
"One suspects that "numerical impartiality" refers to the scrupulously observant use of numerical pronouns, such as "one", where a personal pronoun such as "you" or "I" might be unnecessarily overspecific. If such a suspicion were to bear out, one may reasonably conclude that one's hope to find evidence for the coherent erudition of one's putative better dashed once again."
… which is to say it looks as if he's just talking about saying "one" instead of "you" or "I". If so, he's being as silly as we have guessed.
marie-lucie said,
March 23, 2009 @ 2:51 pm
La vie en rose vs l'avion rose
This may be your favourite pun, but is it a pun in French?
Two reasons why not: first, as mentioned above, the phonemes represented by en and one are not the same; second, vie en counts as two syllables, vion as a single syllable (except in some Southern French, which on the other hand would make more of a difference between en and on than current Northern French).
marie-lucie said,
March 23, 2009 @ 2:53 pm
p.s. the only French person who would misunterpret la vie en rose as l'avion rose would be a small child in Southern France, not familiar with the meaning of voir la vie en rose.
Kris Rhodes said,
March 23, 2009 @ 3:33 pm
Totally sticking my neck out here.
I don't know anything about the guy who y'all are talking about, but reading the three sentences quoted in the original blog post here, I find I can make sense out of them. Whether the sense I've made is the sense intended is another question, but it seems plausible.
"Grammar and logic free language from being at the mercy of the tone of voice."
The ability to respond strictly to denotation plus syntax contributes to an ability to ignore connotative effects. (And sometimes this is good. I've certainly found it constructive sometimes to ignore aspects of the wording of an argument I find personally insulting or whatever, and instead simply to respond to it in terms of its structure and the truth of its sentences strictly interpreted.)
"Grammar protects us against misunderstanding the sound of an uttered name;"
Our ability to interpret utterances as grammatical sometimes helps us disambiguate phonetically unclear sounds. That's pretty uncontroversial, isn't it?
"logic protects us against what we say having double meaning."
The tools of logic allow the construction of formal languages, one defining characteristic of which is the complete absence of semantic ambiguity.
There's certainly something to disagree with in each of these three claims, but they're not in some way utterly ridiculous or implausible as a lot of you seem to think. But maybe I've overinterpreted the sentences or something.
I think one possible problem is that you guys are interpreting "frees us" and "protects us" and phrases like that as meaning "always frees us" and "always protects us." On the other hand, I think he just means "tends to free us" or "allows for our being freed from," and "tends to protect us" or "yields the means for protection from" or words to that effect.
Mark Liberman said,
March 23, 2009 @ 3:51 pm
Kris Rhodes, trying hard to make sense of R-H: "Grammar and logic free language from being at the mercy of the tone of voice." The ability to respond strictly to denotation plus syntax contributes to an ability to ignore connotative effects.
But "tone of voice" doesn't mean "connotative effects", it means, well, "tone of voice". A good actor can infuse an arbitrary word (say, "September") with anger, longing, boredom, uncertainty, or other emotions and attitudes and implications, by varying "tone of voice". But the same is true for arbitrary phrases, sentences or syllogisms.
It seems more likely to me that R-H is trying to say, in a complex and misleading way, something very banal — that referential symbols and compositional syntax and semantics allow us to express things that could never be said merely with growls, moans and whimpers.
marie-lucie said,
March 23, 2009 @ 4:07 pm
myl R-H is trying to say … referential symbols and compositional syntax and semantics allow us to express things that could never be said merely with growls, moans and whimpers
But R-H is not referring to "growls, moans and whimpers", which would not be "language", he probably means strings of independent words which could not be interpreted as being related to each other in a sentence except through "tone of voice" (by which he probably means a combination of pitch, stress and intonation).
Kris Rhodes said,
March 23, 2009 @ 4:20 pm
@Mark
'But "tone of voice" doesn't mean "connotative effects", it means, well, "tone of voice". A good actor can infuse an arbitrary word (say, "September") with anger, longing, boredom, uncertainty, or other emotions and attitudes and implications, by varying "tone of voice". But the same is true for arbitrary phrases, sentences or syllogisms.'
I may have used a technical term incorrectly without realizing it. I would have called the use of the voice to infuse 'September' with anger, longing, or whatever, a "connotative effect." Paying attention strictly to the denotation of "September" and its position in the sentence can help us, if we want to for some reason, to ignore these kinds of effects.
"It seems more likely to me that R-H is trying to say, in a complex and misleading way, something very banal — that referential symbols and compositional syntax and semantics allow us to express things that could never be said merely with growls, moans and whimpers."
Enabling the use of things other than growls to express thoughts is not "freeing language" in any sense I can figure out, since growls are not language.
Taking a look at the quote again:
"Grammar and logic free language from being at the mercy of tone of voice."
I bet you guys are thinking he means that there could be a such thing as a language which is without grammar, and which is not amenable to logical analysis, and that such a language would be "at the mercy of tone of voice." I, on the other hand, am attributing to R-H basic knowledge of the fact that there can't be such a thing as language without grammar–that grammar is constitutive of language. (Maybe I'm wrong here, but we have very little text to go on here, and when in doubt, I like to assume the person I'm interpreting isn't an idiot.) Given that, he can't be saying something that would imply that "language without grammar" is at the mercy of tone of voice.
bianca steele said,
March 23, 2009 @ 4:52 pm
Taking the first sentence alone, without any context, I would have guessed that R-H is simply saying unlearned people are vulnerable to "tone of voice." In other words, he wants to persuade his listeners that it is important to study and to know grammar. I might guess as to why. Presumably, he thinks his audience needs to study grammar and won't think studying is important unless he tells them. Possibly, he thinks his audience is not teaching grammar and needs to be told their students are vulnerable to "tone of voice." Possibly, he thinks his audience will take political action against the forces that motivated their teachers not to teach grammar when they realize they have been left vulnerable. But, taken all together, it makes no sense. It makes you think he uses language for some purpose other than representing meaning that can be understood as such by other humans.
Ray Girvan said,
March 23, 2009 @ 5:25 pm
This may be your favourite pun, but is it a pun in French?
a) Not my favourite pun; b) not strictly – it's more of a cross-lingual joke based on English mangling of French.
marie-lucie said,
March 23, 2009 @ 6:19 pm
"Grammar and logic free language from being at the mercy of tone of voice."
This sentence, although grammatical, cannot be semantically well-formed if different people read very different meanings into it. I see "grammar and logic" (according to R-H) as superimposed on a basic language foundation which must consist of words, not as integral parts of language. Growls and whimpers, unmentioned by R-H, do not enter into it.
peter said,
March 23, 2009 @ 7:34 pm
Even on a blog devoted to language and linguistics, it seems, it is possible to have difficulty parsing an everyday English phrase. Who could have imagined, for example, that one meaning of "main purpose" could be "side effect"?
SJ at 1:02 pm on 23 March:
"I am not suggesting that noun classes arose to increase redundancy. I am merely pointing out that that is one of their side effects."
SJ at 4:32 pm on 22 March:
"With regard to ambiguity (which is a necessary condition for a pun) I would have thought it clear that one of the main purposes of grammatical gender is to provide redundancy."
Kris Rhodes said,
March 23, 2009 @ 7:58 pm
I do not agree with a lot of what Stephen has said in this thread, but its fairly clear to me that he means that increasing redundancy is a side effect of noun classification which has been "selected for" (so to speak) in the "evolution" (so to speak) of certain languages, making it one of the main purposes of that classification.
Much like the aerodynamic side effects of feathers were selected for over time and eventually became one of their main purposes.
peter said,
March 23, 2009 @ 8:50 pm
Kris — One has to be very careful in talking about evolutionary ideas (even metaphorically) if one wants to avoid teleological implications. In the standard contemporary (neo-Darwinian) account of evolution in biology, aspects of living entities that arose through evolution did not do so with a purpose or goal, and, having arisen, still neither have nor exhibit any purpose. Feathers help birds to fly. That is an attribute, a characteristic, a feature, of feathers, not a purpose of them. A purpose implies intention and intentionality, and neither feathers, nor evolution, are entities with intentionality, at least in the standard account. (There are dissenters to this standard account, not only among Creationists and supporters of so-called Intelligent Design. See for example the work of Scott Turner at SUNY Syracuse.)
Insofar as human language arose without a human designer or designers creating it, and insofar as it has subsequently changed without deliberate intent, the same is true. A reduction in the possibility of redundancy may be a feature of noun-class systems in languages (although no one on this thread has yet provided evidence of this claim in my opinion), but it could not possibly be a purpose, since that would imply some entity having a capability for intentionality. Who or what is this entity? It certainly can't be noun classes, or grammar, or a language, since these are all human abstractions (and historically quite recent ones). Could it be perhaps a population of speakers of a language? But then what sense does it make to speak of the intention of a group of people, particularly a group that spanned thousands of years? I believe this makes no sense at all, as silly a notion as Rousseau's "General Will" of Society.
You may object to my argument that I am quibbling over words. Well, words, even (or especially) words describing metaphors, have power over our thoughts, and over what we are able or not able to conceive.
Kris Rhodes said,
March 23, 2009 @ 9:39 pm
Peter,
I used the term "purpose" to mean "function." I should be more careful about this, of course. But AFAIK it is uncontroversial that biological structures have functions. This does not imply that any intention went into the development of their structure.
As it happens, I think a perfectly naturalistic account can be given of biological purposes, one which does not ascribe intentions to their development. But that discussion is far afield from the topic of the present conversation.* I mention it here just to explain why I sometimes use the term "purpose" when most people would think I should say "function" instead.
I did not take Stephen to be ascribing an intention to anyone in the design of languages with noun classes. I took him to mean by "purpose" something more analogous to biological function. The purpose of something, in this sense, is not what some intentional agent designed it to do, but rather, is a set of effects it has which explains its continued existence through a selection process. (Where "selection process" should be read naturalistically, as not implying intention.)
*Oh all right, since you asked. A biological structure functions to do X just in case it was naturally selected for its doing X. Something was naturally selected for its doing X just in case its doing X explains its differential reproductive success. So, for example, a heart has a function–its function (one of them) is to pump blood. What this means is that the heart comes from a line of biological structures–basically, the hearts of the ancestors of the organism it is in–which were selected for their blood-pumping properties. That, in turn, means is that their blood-pumping property explains the fact that their organisms did a better job reproducing than certain other organisms.
See Ruth Millikan, among others.
Nat said,
March 23, 2009 @ 11:12 pm
Like Kris, I didn't see anything terribly remarkable about connecting logic with a lack of ambiguity. The equivocal use of terms can undermine an argument which appears to have a sound logical structure. But it's not so much that logic is protecting from ambiguity as that avoiding ambiguity is a prerequisite for constructing a logical argument.
I'm afraid it's not quite right to say that a logical language avoids all semantic ambiguity. Any set of logical sentences has an infinite number of semantic interpretations. But a logical translation of a natural language sentence, i.e. a logical statement plus an interpretation, does avoid ambiguity.
Stephen Jones said,
March 24, 2009 @ 6:49 am
The standard example of a biological organs present function being different from its original function is the bird's wing. Nobody would deny now that its primary function or purpose is to enable flight. But that could never have been its original function since for the first few hundred generations of its development it would not have been able to allow flight and must have had some other purpose (improving balance while running being the most likely).
Kris has understood what I'm saying perfectly.
peter said,
March 24, 2009 @ 7:03 am
Many commenters here seem to believe that ambiguity, imprecision, illogic and misunderstanding in communications are always undesirable features. Can I just say again that this is not necessarily the case. Many communicators (both speakers and writers), on many occasions, desire just these characteristics, and they are not irrational to do so. Puns are a common example, but there are many others more serious than this. Much of nternational diplomacy and treaty negotiation, for example, could be seen as aiming at the elegant expression of mutually-incompatible ideas in single statements.
acilius said,
March 24, 2009 @ 8:28 am
@Kris & Peter: You two are getting very interesting. I hope you keep this discussion going for a bit.
Kris Rhodes said,
March 24, 2009 @ 9:11 am
@Nat "I'm afraid it's not quite right to say that a logical language avoids all semantic ambiguity. Any set of logical sentences has an infinite number of semantic interpretations."
Yes, that's right. And checking, I see that people knew this (it's Lowenheim-Skolem that's relevant, right?) way back in 1915. I didn't know it was so early.
I've always wondered something about the implications of that theorem. It says all sets of logical sentences have a countable infinity of models. (Right?) But doesn't the specification of a formal language also include a specification of the semantics of it? Doesn't that determine a unique model? Or is the specification of semantics considered to consist in a further set of logical sentences? (I don't think so but it's easy for me to be wrong about this kind of thing.) Now, I know that in practical terms, there's bound to be ambiguity in the specification of the semantics. But is that, in itself, a result of the truth of L-S, or is it a separate fact?
Anyway I was just trying to interpret H-R's sentence, and I think it's plausible that a philosopher of H-R's school wouldn't have known about this result. Before learning about the technicalities, I think people generally did (and do) think formal languages lack ambiguity.
Kris Rhodes said,
March 24, 2009 @ 9:14 am
"Many commenters here seem to believe that ambiguity, imprecision, illogic and misunderstanding in communications are always undesirable features."
I may be mistaken, but I think you're aiming this comment at a group of people that includes me. But if that's the case, then you're mistaken. I think H-R dislikes ambiguity, imprecision, and so on, but I myself have not expressed any of my own feelings on the subject. I agree with you that they are not always undesirable features.
"Much of nternational diplomacy and treaty negotiation, for example, could be seen as aiming at the elegant expression of mutually-incompatible ideas in single statements."
That, however, I don't see the value of. I like to be educated, though.
Kris Rhodes said,
March 24, 2009 @ 9:16 am
@Nat: I wish I could edit posts here. Somehow I missed the last sentence in your post, so my question about interpretations should be ignored. Sorry about that!
Robert said,
March 24, 2009 @ 10:53 am
Technical correction: The Loewenheim-Skolem theorem applies to countable theories using first order logic only. Counterexamples exist to a stronger statement, for instance the second order theory of real-closed fields has only one model (perhaps it's unique up to a unique isomorphism instead of actually unique, but the point is you don't get models of every infinite cardinality like Loewenheim-Skolem).
Sridhar Ramesh said,
March 24, 2009 @ 2:41 pm
@Robert: It sounds like you probably know this, but just to put it out there, versions of Lowenheim-Skolem apply to any first-order theory, whether or not the theory is countable. But you are correct that a unique model (up to isomorphism) can be achieved with a theory in second-order logic, or other such logics which lack a compactness theorem; an even simpler example than the one you give is the second-order Peano Axioms, whose only model is the standard natural numbers.
@Kris Rhodes: "But doesn't the specification of a formal language also include a specification of the semantics of it?". If I understand what you mean by this question, the answer is, in a word, no; at least, in the sense of the theorem you are referring to, an interpretation of a language is allowed to be basically any mapping from the sort, function, and relation symbols in that language to actual domains of discourse and functions and relations upon them (and a model of some collection of sentences is allowed to be any interpretation which makes them all come out true).
Sridhar Ramesh said,
March 24, 2009 @ 3:08 pm
(I guess I should also pre-emptively say, there are two senses in which "second-order logic" can be used; there's the one Robert and I used above, so-called "standard semantics", in which quantification over "subsets of X" is always interpreted as quantifying over the actual powerset of X; however, there is also "Henkin semantics", in which the domain serving as the interpretation of "subsets of X" need not actually contain all the subsets of X. Less sentences are semantically tautologous with Henkin semantics (due to a wider class of models being considered), but it does recover all the same basic properties as first-order logic enjoys, including the Lowenheim-Skolem properties)
Nat said,
March 25, 2009 @ 2:33 pm
Kris, I was actually thinking of something a bit simpler than L-S, namely that the interpretation (the mapping from names, predicates, function symbols to objects, sets/properties, functions) needs to be specified in addition to the language if you want to say something definite. That's just in the nature of using of logic. So, awareness of this is going to go back to Aristotle. The whole point of syllogisms is that each one has multiple models (so, can represent multiple arguments).
Also, those "up to isomorphism" clauses are important. Talking about the same thing up to isomorphism might be fine in a mathematical context, but it's not satisfactory for daily life. If I want to know what to do in an emergency, it's not helpful to be told to dial 9-1-1, when "dial" can refer to any predicate, and "9" and "1" can refer to any two objects.
norman fiering said,
March 31, 2009 @ 10:36 am
Would anyone seriously propose that the worth of a book, or even an essay, can be judged on the basis of a few sentences from it? Among the 500 or so publications listed in Lise van der Molen's GUIDE TO THE WORKS OF EUGEN ROSENSTOCK-HUESSY (1997) there is MAGNA CARTA LATINA (2nd ed., 1975), which is a Latin grammar book based entirely on sources from between 1000 and 1500, an anti-Humanist tract, of sorts. Here is a relevant quotation from p. 15: "When a man writes a patriotic poem or a whole book, he tries to communicate a responsible idea to responsible people. His poem is one whole. His book is a unity. You cannot break it up piecemeal without destroying its meaning. Articulated speech, then, may be as long as the Bible and still convey only one idea; in the case of the Bible, that of the government of the living God. The whole Bible, then, is a unity out of thousands of sentences and tens of thousands of words. . . . A brick bridge and heap of bricks do not have the same meaning. . . . We cannot expain the bridge by explaining the bricks."