Lambda the Ultimate

What does that make Java?

Though this article makes for entertaining reading, it makes a number of points that make good stories but lousy linguistics.

In particular, the bit about "evidentiality" is a classic case of exoticism: the article makes it sound that because there is a morphological rather than syntactic way of expressing this, that this must result in a different mentality about sources of information.

This is precisely the equivalent to saying that because Latin can express the future tense with conjugation rather than using a separate word "will" as in English, the Romans must have had a deeper concern with the future.

Given the sentence "Unicorns are white.", I can easily say: "They say unicorns." "I can see that unicorns are white." "I've heard unicorns are white." "I believe unicorns are white."

And so on. Each sentence will give you a different expectation about my truthfulness, my knowledge, and possibly my sanity. ;-)

To draw this back to CLs, let's consider the property of "foldability": particular FP has a predefined "fold" operator.

Now if I write a program in that language, but don't use the fold operator, I build my own folds from scratch instead, "native speakers" of the language will think I am an unreliable programmer or a beginner or a "foreigner" (adherent of some other language), but is this really a comment on the properties of the language itself, or a comment on the prevaling fashion and social rules that govern those circles?

I agree with your general skepticism, but I think that it is far from clear just how much, and exactly when, differences between languages influence the ideas being expressed. This is linguistic relativity all over again. I tried to debunk some of the myths about this in earlier threads, but I continue to claim that simply dismissing this notion is also hard to support.

As to evidentiality: If you must include the relavent suffix in language A, where as in language B you may choose to provide this information or not, then it stands to reason that the language can have some influence in this case. It doesn't mean that A is more expressive (whatever that means) but that's not the only aspect worth exploring.

<>

Babel's Children, The Economist, January 10th-16th 2004

You can easily express contracts a la Eifell using regular if statements found in other languages. Is this enough to conclude DbC is of no linguistic importance?

<http://itre.cis.upenn.edu/~myl/languagelog/archives/000128.html>

Someone explain why one statement is proverbial and the other comical:

"French, the language of love."
"German, the language of love."

First of all, I definitely think that both evidentiality and DbC are of significant linguistic importance. I am simply claiming that they are not important in the way that is implied by the article.

In the former case, if you are a linguist studying the language in question, or someone trying to learn the language to be understood by native speakers, then it is vital that you understand how to use the mechanism of evidentiality.

However, to suggest that this mechanism introduces a "foreign concept" to a language that lacks it, or that it is harder to dissemble in a language with this feature is silly.

In the DbC case, I think the liguistic analogy is a poet's use of language. DbC as a concept exists independently of the language, as does moonlight shining through the trees, but if Eiffel were for poetry, it would give the poet a convenient word that scans for the DbC concepts.

These details are important to the aesthetic experience of using either type of language, but they don't revolutionize the basic semantic possibilities of either.

Simple: we have different stereotypes for the cultures of the respective speech groups.

I have met people (especially from non-English-speaking countries) who thought German was beautiful and French was harsh and staccato.

Here's a classic example of that I learned long ago ("The little birds sing in the trees"):

French: Les petits oiseaux chantent dans les arbres.

German: Die kleine Vögeln singen im Baümen.

To hear the difference, you'll have to ask a native speaker of each language, though!

This reminds me of a short story by Jorge Luis Borges, Tlön, Uqbar, Orbis Tertius. In this story, Borges mentions two languages without nouns, one using adjectives and the other using verbs. For example: "Upward behind the onstreaming it mooned" for "The moon rose above the river". He also has an interesting discussion on how language creates reality. In one language, any arbitrary conjunction of words defines a new concept. In another, reality consists of many small surprises, since the language does not support inductive reasoning (it does not allow for "the coin I found today is the same one that I lost yesterday"). The small surprises are a fact of life; the found objects are called hronir and they are considered to be created by the universe.

If you have not read this story, I highly recommend it. You can find the full text easily through a search engine.

Simple: we have different stereotypes for the cultures of the respective speech groups.

Easy to say; hard to prove. The language stands on its own, like a piece of music. French is spoken in African and Carribean nations by different cultures altogether.

Not that I take any position on the proverb's veracity. I merely observe its existence. The Germanized version did once buy me a hearty round of laughter, from German Lutherans no less, practicing a Bach chorale. (By the way, be careful about assigning or assuming beliefs in cultural stereotypes.)

I have met people ... who thought German was beautiful and French was harsh and staccato.

Having learned both languages I might quibble, but the question is whence came the proverb.

To hear the difference, you'll have to ask a native speaker of each language, though!

I guess so, because I discern no major lesson in contrast here.

This reminds me of a short story by Jorge Luis Borges...

Ah. Be careful about Borges. When I met a literature expert familiar with him, I seized the chance to ask, in reference to stories like this one, "What in the world is all that stuff about." He smiled and explained his reading of Borges as a man fond of exploring the possibilities of fiction. For him there was no deeper meaning than that. Anyway I'm not sure the story supports Marc, if its lesson is that "language creates reality."

Actually Marc, I'm no linguist, and we have no real argument. After all there is a difference between talking to fellow humans and talking to computer chips. I imagine that each task exercises different realms of the human brain. So it's possible that some future understanding of linguistic issues down to the last brain cell will leave various PL issues unresolved. PL issues follow rules of their own. Consequently one might agree with you on all matters linguistic and disagree on PLs, with perfect validity. I think you move too quickly from Unicorns to fold operators. What is untrue of human languages might still be true of PLs.

...is this really a comment on the properties of the language itself, or a comment on the prevaling fashion and social rules that govern those circles?

Consider George Orwell's thesis on the deliberate dumbing down of human language as a technique to limit human thought. There is a reason to value high vocabulary in both human and programming languages. It boils down to economy, precision, and accuracy of thought, weightier matters than "aesthetic experience." Any politician will tell you that framing a debate in his chosen terms is next of kin to victory. In the PL world "victory" might be defined as O(N log N) instead of O(N^3).

PVR's book states that concurrency is considered difficult by Java programmers simply because of language support issues. They cannot think concurrently because they lack the Java vocabulary for the concept. If they spoke Oz, then they would not force-fit every problem into single-threaded solutions. (Forgive my paraphrasing butchery, PVR.) Some of the ASoC papers referenced in my Sina post delve into the problems of object composition in deficient languges, e.g. "The [proposed] model ... allows the programmer to express a seamless spectrum of composition semantics in the interval between object composition and inheritance." The offering is a spectrum of expression, not just "good," "plus good," and "double plus good." I hope that post generates its own discussion.

Perhaps my familiarity with both languages is interfering, but I don't notice any significant aesthetic difference between the two.

Oh well, I guess that goes along with my point. ;-)

Let me provide a counterexample based on this:

Senegalese French is the language of love. Viennese German is the language of love.

This pair is more likely to produce bewilderment(or poetic travel writing ;-) than humour.

a hearty round of laughter, from German Lutherans ... (By the way, be careful about assigning or assuming beliefs in cultural stereotypes.)

This just shows that the Germans have a similar stereotype of themselves. ;-)

What is untrue of human languages might still be true of PLs.

I would put this much more strongly: there definitely features of PLs that are entirely unlike human languages.

A major example: there exists a widely-agreed-upon mechanism for defining the semantics of a PL. No such thing in HL (nor likely to be anytime soon).

I would go so far as to say that the difference in the scope of the respective semantics of PL and HL might even account for the relatively greater impact that syntax has on PLs.

Human semantics are so complex that no variations in syntax really have significant effect on the ability of a speaker of any language to cope with a given subject area.

On the other hand, "syntactic sugar" that is built up from a well-defined primitive semantics of PLs seems to make the difference between a particular type of programming being straight-forward or not. (Think using closures in assembly)

Consider George Orwell's thesis on the deliberate dumbing down of human language as a technique to limit human thought.

Unfortunately, I think this is a case of Orwell making a moral issue out of his own cultural and class preferences (as well as a misunderstanding what is happening to English as it changes). I would expand, but I don't want to go to far off topic. ;-)

As a German-American Lutheran, I can't resist pointing out that this is very nearly a tautology.

This pair is more likely to produce bewilderment ... than humour.

Which fact makes my point. Explicit reference to culture groups in a parable about language produces bewilderment. When I hear "French is the language of..." what comes to mind is French qua French, not Frenchpeople.

This just shows that the Germans have a similar stereotype of themselves. ;-)

It shows, rather, that they know their language only too well.

I would put this much more strongly: there definitely features of PLs that are entirely unlike human languages.

All right then: even if Orwell is wrong about HLs, his principle might hold true of PLs, because we agree that HLs and PLs are worlds apart.

Paul: ...this is very nearly a tautology.

Not quite, owing to a large population of German Catholics!

French: Les petits oiseaux chantent dans les arbres.

German: Die kleine Vögeln singen im Bäumen.

The French sentence has 'ch' and 'rbr' and 'z - z' (soiseaux). The German is much smoother: 's' of singen is pronounced 'z', lots of 'n' and 'm'. I agree the difference is not world-shaking, but it's there and it goes against the usual stereotype of German being full of harsh sounds and French being smooth.

(In the area where I live, Tyrol, it could be spoken actually (with pseudo-english phonetics), Dee kloanen Foihgel singen in dee bah-mah.*)

More common than singen for the sounds of birds is zwitschern which probably fits the stereotype well.

* At least in my part of the Inn-valley. :-)

The unspoken assumption here, and in the converse of this remark, is that something that is harsh and/or staccato cannot be pleasing. However, that is clearly an aesthetic judgement.

I enjoy gothic rock and industrial music, and a lot of it is in German; the amount of such music in French is, well, negligible. Granted most people hate these genres, and goths have an <cough> unusual sense of aesthetics, but some of us (and I know of one or two who frequent this very forum) like to think it reflects a deeper sort of capacity for appreciating things.

OTOH, I have to admit that I really hate the more extreme subgenres like (power)noize. :)

[T]here definitely features of PLs that are entirely unlike human languages. A major example: there exists a widely-agreed-upon mechanism for defining the semantics of a PL. No such thing in HL (nor likely to be anytime soon).

I quote Marc's earlier remark:

[I]s this really a comment on the properties of [human] language itself, or a comment on the prevaling fashion and social rules that govern those circles?

Thanks for improving on my Schuledeutsch!

The purpose of Borges stories, as I understand it, is to explore ideas. They're usually quite well thought out. For example, another particularly interesting one is Funes the Memorious, exploring the inner thoughts of a man who literally cannot forget anything.

It seems that Borges wrote many more essays than short stories, but sadly only a few are translated into English. I have found just one book of such essays, and I enjoyed it immensely. I wish I could read Spanish as well as English, just so that I could read more Borges essays.

I'm not sure what you mean Frank: the remarks seem utterly orthogonal to me .

One says HL and PL are very different in the scope of their semantics. The other says that HL and PL are similar in their sociological properties.

Definitely so.

Look at the popularity of Klingon for example. ;-)

My point was that the existence (or lack thereof) of "a widely-agreed-upon mechanism for defining the semantics" of X is less "a comment on the properties of [X] itself" than "a comment on the prevaling fashion and social rules that govern [X's] circles."

Oy, not more human factors. ;-)

OK, so let's back up and see if I can provide a more satisfactory statement of position.

I want to provide some evidence that HL semantics are significantly more complex than PL (or, more generally, computational) semantics.

The best I can do is to note that "great minds" have been working on general semantics for centuries, and haven't yet come up with a compelling and comprehensive theory that works.

In computation, by contrast, several people came up with simple formalisms that clearly work (and provably so) independently and at the very birth (or even before the birth) of the field.

True, philosophers of language may be dumber or more disagreeable than computational mathematicians, which would provide a purely sociological explanation for the disparity, but I feel pretty good about saying that HL semantics is just way harder. ;-)