Two brews

« previous post | next post »

The cover of the January-February issue of the Pennsylvania Gazette has a nice picture of Victor Mair lifting a glass with Patrick McGovern, to illustrate an article about Patrick's new book Uncorking the Past: The Quest for Wine, Beer, and Other Alcoholic Beverages, and an interview with Victor about his new book with Erling Hoh, The True History of Tea. (There are some excerpts from the two books as well.)

The reasons to be interested in the books, and to enjoy these two articles in one of the best alumni magazines around, are mostly not linguistic ones.  However, a couple of language-related points stuck with me.

One was Patrick McGovern's title, "scientific director of the Biomolecular Archaeology Laboratory for Cuisine, Fermented Beverages, and Health at the Penn Museum".   My reasons for finding this interesting are non-obvious, not to say weird.

I've noticed over the years that a surprisingly large fraction of smart undergraduate students have a surprising amount of trouble with what seems to me like a spectacularly simple-minded idea: the simple parallelism between form and meaning that linguists generally call "recursive compositionality", and compiler writers  call "syntax-directed translation".

A trivial example of this would be the relation between form and meaning in arithmetic expressions: thus in evaluating (3+2)*5, you first add 3 and 2, and then multiply the result by 5; whereas in evaluating 3+(2*5), you first multiple 2 and 5, and then add 3 to the result.  Similarly, in evaluating English complex nominals, the phrase stone traffic barrier normally means a traffic barrier made out of stone, not a barrier for stone traffic, and thus its meaning implies the structure (stone (traffic barrier)).  A plausible way to think about this is that you first create the phrase traffic barrier, and the associated concept, and then combine that — structurally and semantically — with stone.  In contrast, the phrase steel bar prices would most plausibly refer to the prices of steel bars, and thus implies the structure ((steel bar) prices).

There are many formalisms for representing and relating linguistic form and meaning, but all of them involve some variant of this principle.  It seems to me that understanding this simple idea is a necessary pre-condition for being able to do any sort of linguistic analysis above the level of morphemes and words.  But when I first started teaching undergraduate linguistics, I learned that just explaining the idea in a lecture is not nearly enough.  Without practice and feedback, a third to a half of the class will miss a generously-graded exam question requiring them to use parentheses, brackets, or trees to indicate the structure of a simple phrase like "State Department Public Relations Director".

In fact, even a homework assignment with feedback is not always enough, even with a warning that the next exam will include such a question.  For some reason that I don't understand, this simple analytic idea is surprisingly hard for some people to grasp.

Anyhow, I generally create a homework assignment and a couple of exam questions to train and test these concepts and skills. And in order to prevent some students from relying on the archives of past homeworks and exams stored (I'm told) at some fraternities and sororities, I need to find new examples every year. So things like "scientific director of the Biomolecular Archaeology Laboratory for Cuisine, Fermented Beverages, and Health at the Penn Museum" go into my file of possible future homework questions.

And in a completely different vein,Trey Popp's article on Patrick McGovern also has a nice example of parodied wine talk:

“We think that chewing is probably the earliest way that humans would have transformed starch into sugar” in order to get fermentation going, McGovern explained in his lecture. Modern beer relies on the enzymes released by sprouted and malted barley to get this job done, but malting barley can be a little tricky. Our saliva contains an enzyme called ptyalin that does the same thing. “It may not sound very appetizing to think of people preparing their beverages this way, but once you get an alcoholic beverage, it does kill off any harmful bacteria.”

He paused for a beat. “And it might add some special flavors, too. You never know.”

Well, not until you slug it down, anyway. Compared to the saffron-kissed honey of Dogfish Head’s Midas Touch, or the chili-tinged chocolate notes of its Maya- and Aztec-inspired Theobroma, the special flavors in the spit-primed Chicha ran more toward funky pink peppercorns with a hint of fraternity basement.


  1. Jenno said,

    February 6, 2010 @ 1:44 pm

    A fun trick question for your exam would be "Department of Redundancy Department."

  2. Jonathan Badger said,

    February 6, 2010 @ 1:59 pm

    In contrast, the phrase steel bar prices refers to the prices of steel bars, and thus implies the structure ((steel bar) prices).

    But in most languages the issue is that there isn't an unambiguous interpretation of such phrases, and the correct interpretation is simply up to plausibility and context. The classic example that proponents of logic-based conlangs like to use is "pretty little girls' school", which can have a number of meanings in English. ((pretty little girls') school), (pretty (little girls' school)), ((pretty little) girls' school)) and so on.

    [(myl) Exactly. You understand the relevant aspects of the relation between form and meaning, you know how to symbolize them using parentheses, and you even understand how to explain the concept of structural ambiguity in terms of such representations. Full marks for you!]

  3. John Lawler said,

    February 6, 2010 @ 2:29 pm

    Not just in syntax, of course. Morphology works the same way (though English examples are hard to find). My favorite is an ambiguous one: unlockable, which can be parsed either as (un-(lock-able)) 'not capable of being locked', or as ((un-lock)-able) 'capable of being unlocked'.

  4. marie-lucie said,

    February 6, 2010 @ 2:29 pm

    In Canada, just about anything sold for public consumption (I don't mean intellectual consumption of books, etc) has to have its descriptive material printed in both official languages, English and French. It is very rare that material in French needs to be translated into English, so the majority of translations are from English into French, and complex English noun phrases often pose a problem, as in "Wild Forest Mushroom soup – soupe aux champignons de la forêt sauvage" (soup with mushrooms from the wild forest). The versatility of English words which can be used as more than one category is an additional challenge: just a few days ago I saw an ad (I think it was for pillow cases or other bedding material) which claimed the following: "Mildew proof – Preuve de Moisissure" (proof (N) of mildew). (I must say that such translations have improved in the decades I have lived in Canada, but some manufacturers – especially if not Canadian – don't seem to bother with hiring professional translators).

  5. Christopher Henrich said,

    February 6, 2010 @ 2:38 pm

    I am a computer programmer, not a linguist. It seems to me that the analysis of a "complex nominal" like (stone (traffic barrier)) goes beyond what I think of as "syntax-directed translation."

    In examples of syntax-directed translation, such as for 3 + 4 * 5, the compiler relies on fixed rules such as precedence: * binds more closely than +. It gets more complicated if the operators are "overloaded" with different meanings depending on the types of the operands. This a + b*c might be translated one way if a, b, and c are integers and quite another way if a and c are vectors while b is a matrix. This is "syntax-directed" if we allow "syntax" to include looking up the declared types of identifiers.

    Computer scientists have sometimes amused themselves by concocting programming languages that are mind-bendingly hard to use (see INTERCAL). I don't remember any in which the phrase structure (such as precedence of operators) is different for different types of operands. (Oh God. Somebody will take this statement as a challenge. What have I done??)

    And yet it appears that English already is such a language. We accept (stone (traffic barriers)) but boggle at ((stone traffic) barriers) because (stone traffic) is implausible. On the other hand, ((wheeled traffic) barriers) is preferable to (wheeled (traffic barriers)).

    I think I recall someone (a linguist) saying that semantics does not influence parsing. Really?

    [(myl) Computer (and other formal) languages are constructed so as to avoid ambiguity, or to resolve it by general principle (e.g. operator strength and default left-branching structure in arithmetic expressions. Natural languages, in contrast, appear to allow and even revel in structural ambiguity (as well as other kinds). So applying the analogous ideas about structure-guided interpretation, you need to recognize that there's often more than one way to do it.

    And everyone in the field understands that most English phrases are structurally ambiguous, and that the ambiguities are resolved (to the extent that they are) by semantic and pragmatic means. There are disagreements about how this happens (e.g. by a process of selection among alternative (complete or partial) analyses, or by a process that integrates information in a more bottom-up way so that the "wrong" analysis are never really considered at all), but no disagreement about the overall process.

    You could (and some people have tried to) construct a theory in which there are no real structures at all except for the patterns (e.g. of neural net activation) that underlie complex meanings. But nearly all respectable analytic frameworks have some notion corresponding to the form-meaning relationship described above.]

  6. John Roth said,

    February 6, 2010 @ 2:57 pm

    I think I recall someone (a linguist) saying that semantics does not influence parsing. Really?

    "Pretty little girl's house" is an already mentioned counterexample, where the context is necessary to determine the proper parse.

    This attitude is probably why it's possible to create natural language interfaces for single domains, but it's proved very difficult to create a single interface for multiple domains. It doesn't seem like parsing English should be that difficult, as long as it's done to add to an existing pragmatic model (is that the right term?) rather than in splendid isolation.

  7. Amy Stoller said,

    February 6, 2010 @ 3:14 pm

    "It's a naive domestic burgundy without any breeding, but I think you'll be amused by its presumption" has become a family saying. Gold star for the first correct identification of origin.

  8. Sam said,

    February 6, 2010 @ 3:30 pm

    OK, so on your exam, which do you accept: ( (State Department) ((Public Relations) Director) ) or ( ( (State Department) (Public Relations) ) Director ), and do you have a reason for preferring one over the other?

    It seems to me that either analysis can be valid, depending on whether you're emphasizing a role as Director of the State Department section of Public Relations, or a role as Director of Public Relations for the State Department.

    [(myl) Either answer is fine; extra credit for pointing out the ambiguity and characterizing it correctly.]

  9. Simon Cauchi said,

    February 6, 2010 @ 3:58 pm

    May I quibble about your first sentence? It's Patrick McGovern who's raising a glass. Victor Mair holds a tea cup.

  10. Charles said,

    February 6, 2010 @ 4:14 pm

    Amy Stoller: Thurber, of course.

  11. Aaron Toivo said,

    February 6, 2010 @ 4:28 pm

    English examples with tree-like morphological structure aren't that hard to find. One beauty is that infamous pseudo-word, antidisestablishmentarianism. It's good for showing how every layer of morphology creates a well-formed word, the whole way through:


  12. Adrian Bailey (UK) said,

    February 6, 2010 @ 4:30 pm

    Marie-Lucie, part of the problem is that (and this is connected with the difficulty Mark has with his students) too few English-speakers know how to hyphenate. For example, if the original texts had been "mildew-proof" and "wild forest-mushroom soup", the translations would probably have been correct.

    (Okay, the hyphenation in the soup example is optional, but I'd recommend it for the sake of understanding.)

  13. Amy Stoller said,

    February 6, 2010 @ 4:46 pm

    @Charles: Right first time! [Exit, pursued by a barking seal.]

  14. Darin Ramsey said,

    February 6, 2010 @ 5:09 pm

    Thank you so much for the mathematical method of explaining complex nominals. I edit for engineers, for whom six consecutive descriptive nouns seem to be taken as a minimum (e.g., "pulse jet mixer vent filter system"). This parenthetical method will help me explain to them when they need to hyphenate, for example, or when they need to rephrase to "vent filter system for the pulse jet mixers." (Getting them to write "pulsed-jet mixers" is, alas, against the project's style guide.)

  15. Simon Cauchi said,

    February 6, 2010 @ 5:29 pm

    @Darin Ramsey: Shouldn't that be "filter system for the pulse jet mixer vents"?

  16. John Cowan said,

    February 6, 2010 @ 5:32 pm

    Aaron Toivo: Nothing pseudo about the word anti-disestablishmentarianism: it's the political beliefs of those who were (successfully) opposed to disestablishing the Church of England (that is, making it cease to be the established church), or alternatively of those who were (unsuccessfully) opposed to disestablishing the Church in Wales.

    Why in Wales as against of England? A parsing error. The Act of Parliament which disestablished the Welsh branch of the Church of England referred repeatedly to "the properties of the Church in Wales", meaning the church buildings and lands and what not belonging to the Church of England but physically within the boundaries of the Principality. The resulting separated organization which took ownership of those properties believed that by statute its name was "The Church in Wales", due to the misparse of "properties (of the Church) (in Wales)" as "properties (of the Church (in Wales))". And so it remains.

  17. marie-lucie said,

    February 6, 2010 @ 5:38 pm

    I think that hyphens are going out of style. For instance, many compound verbs (a relatively recent phenomenon – not a "recency illusion") are not noticed in writing because they are just not hyphenated. There are a few (very few) older ones which are usually written as one word, like to babysit, but the new ones that people are constantly creating are left as two words, as in Do you know how to African dance? or I like to people watch (actual quotes from ads or the media).

    Even in French, I have seen for instance dit-on or donne-le-moi written separately: dit on, donne le moi.

  18. Rubrick said,

    February 6, 2010 @ 6:18 pm

    Instead of finding new examples each year, why not just vary the number of "buffalo"s? :-)

  19. Karen said,

    February 6, 2010 @ 6:23 pm

    Oh, I don't know. I was taught that hyphenating such things was the midway point, not the beginning – that is, "African-dance" would come along naturally after some time with "African dance", and before "Africandance", as in "babysit" which began as "baby sit", not "baby-sit".

  20. po said,

    February 6, 2010 @ 6:48 pm

    Maybe those undergrad students were confused by the conflicting use of parens in writing? They're used to enclosing sentence parts that can be omitted in parens. You ask them to use the self-same parens for grouping ⇒ hed asplodes.

  21. DaveK said,

    February 6, 2010 @ 7:18 pm

    "Scientific director of the Biomolecular Archaeology Laboratory for Cuisine, Fermented Beverages, and Health at the Penn Museum"
    Speaking of ambiguity, doesn't that comma after "beverages" change the meaning into {(scientific director) of the [(Biomolecular Archaeology) Laboratory] for [(Cuisine,) (Fermented Beverages), (and Health at the Penn Museum)]} instead of {(scientific director) of the [(Biomolecular Archaeology) Laboratory] for [(Cuisine), (Fermented Beverages), (and Health) at the Penn Museum]} ?

  22. Skullturf Q. Beavispants said,

    February 6, 2010 @ 7:19 pm

    I teach college-level mathematics, and when I read about your puzzlement at students' difficulty with analyzing linguistic form using tools like parentheses, I'm reminded of some difficulties I notice in some of my own first-year students.

    This is Language Log, not Math Log, so I'll try to be brief. In calculus and precalculus, we might study functions; for example, the function defined by f(x) = x^2. Then what is f(x+h)? I've noticed surprisingly many students having problems understanding why it's (x+h)^2 rather than, say, x^2 + h, or x + h^2.

    Part of the reason this puzzles me is that it's not difficult in the same way that other things in math can be difficult. It's not difficult in the way that lengthy computations are difficult, or theorems with a long list of conditions are difficult, or elaborate and byzantine proofs are difficult.

    I guess part of it is that students have trouble with things that are more abstract than concrete, even if such things aren't long or technical and don't have an enormous number of details to get lost in.

  23. Skullturf Q. Beavispants said,

    February 6, 2010 @ 7:38 pm

    You almost want to say, "Geez, people, just do things one step at a time, and pay attention to the order."

  24. Dan T. said,

    February 6, 2010 @ 8:12 pm

    The Onion's "Our Dumb Century" book, with fake newspaper front pages from throughout the 20th century, often adds spurious hyphens to create words like "news-paper" in order to lend an "archaic" feel to the earlier-dated pages.

    I can recall from looking at genuine (non-parody) newspaper pages from dates around 1900 that you often found things like "Base Ball" and "Foot Ball"; also, you sometimes saw New York (city and state) referred to as "New-York" with a hyphen. "Pittsburg" was missing its final "h".

  25. Alex said,

    February 6, 2010 @ 9:28 pm

    >>>>I can recall from looking at genuine (non-parody) newspaper pages from dates around 1900 that you often found things like "Base Ball" and "Foot Ball"

    I think that really shows you the power that Microsoft's spell check has over contemporary linguistics. It's hard to tell how many people might have made the same constructions today, because their compositions would have been highlighted in red, imploring them to change their mind on the spelling.

  26. Jerry Friedman said,

    February 6, 2010 @ 9:44 pm

    @Karen: I suspect a lot of people were taught that, unfortunately. It's true for some things—Dan T. just mentioned base ball > base-ball > baseball, which I suspect is how it happened. But the majority of the Google Books hits before 1950 on "to baby sit" baby-sit used a hyphen.

    The hyphen means something: "baby-sit" or "babysit" is a single word, a verb, not a noun plus a verb. Thus you need a log-in (verb plus preposition-or-whatever-you-call-it acting as a noun) so you can log in (ordinary phrasal verb), you eat wild-mushroom soup (noun phrase acting as an attributive noun) made of wild mushrooms (ordinary noun phrase). (On the other hand, I feel that in "Google Books hits", the capitalization removes the need for a hyphen. Purists may disagree.)

    But though the trend may once have been toward writing solid, I think it's now the other way, as Marie-Lucie said. My students almost never use hyphens and many of them write most compounds as two words. You'll have no trouble find hits for "my birth day", though "my birthday" still outnumbers it greatly, I'm glad to say. Unfortunately, this means readers have to figure out the syntax of compound words from the meaning, as people have said above, which may be a little extra difficulty, or a lot if readers aren't familiar with the subject.

  27. speedwell said,

    February 7, 2010 @ 1:58 am

    [(myl) Either answer is fine; extra credit for pointing out the ambiguity and characterizing it correctly.]

    Oh, I don't know. I think "I'm the Director of Public Relations at the State Department" ( (State Department) ((Public Relations) Director) ) is significantly preferable to "I'm the Director of State Department Public Relations" ( ( (State Department) (Public Relations) ) Director ).

    Then again, our Middle East Human Resources Director coincidentally happens to be a human from the Middle East, but anyway… he is the director of the Middle East Human Resources department of our company, not the Director of Human Resources for the geographical area in toto.

  28. Zora said,

    February 7, 2010 @ 3:12 am

    Kava, a South Pacific drink, is prepared from shavings of the root of the Piper methysticum, which are chewed, spit into a bowl, and then mixed with water.

    Or that is how it was traditionally done. Westerners have taught Fijians and Tongans that this is disgusting, and now the ground-up root is merely mixed with water. Probably not as potent without the pytalin in the spit.

    BTW, it tastes like drinking spicy sawdust.

  29. unekdoud said,

    February 7, 2010 @ 6:11 am

    Bonus homework question type: Explain the crash blossom phenomenon in terms of the parenthesis structure given above!

    About use of context in affecting interpretation: Sometimes, the unsaid words have to be taken into account. Take speedwell's example(2 comments up), where leaving out the "our company" part of the description can lead to slightly different readings.

    About readings: usually the interpretation/grouping can be clarified by using a different tone of voice. Not applicable for written/typed text!

    About grouping: How do you use parentheses to indicate that two non-adjacent words are interacting? For instance, we currently join adjacent words to form noun and verb phrases, but what if the connection between two words is indicated by sentence structure?

    About math, and "doing things in order": Part of the confusion arises because of the different ways of composing the two operations. In a linguistic context, this corresponds directly to the existence of different constructions using the same words (in the same parts of speech). It's similar to writing adjectives in the wrong order. This also means that writing according to the order could lead bad ordering of adjectives. We don't say "blue dark jeans" when we want to emphasize the darkness of the color.
    On the other hand, we also have a sort of associativity problem here (if put in math terms): We may want "dark blue" to be an adjectival phrase by itself, but have no syntactical way of forcing that interpretation. In math, an extra pair of brackets would(might) do the trick, but English doesn't have such punctuation – at least, none that are grammatically acceptable and yet natural. Multiple spacing, as well as commas, comes close.

    In fact, the de-spacing of compound nouns is another solution to this problem, but certainly not satisfactory for those 7-word nouns mentioned in the article. Well, the human mind can only take so many levels of nesting and branching, right?

  30. Jarek Weckwerth said,

    February 7, 2010 @ 8:55 am

    unekdoud: About readings: usually the interpretation/grouping can be clarified by using a different tone of voice. Not applicable for written/typed text!

    More technically, this is done using intonation, not "tone of voice". Even more technically, division into intonation phrases, placement of the nucleus and selection of tone. My standard example is this:

    six plus four divided by two equals… how much?

    This is not parsable in writing (the numbers are spelled out on purpose, of course, to preclude the use of the conventions of mathematical notation). In speech, the parse is usually there:

    six plus /four | divided by /two | equals \five
    /six | plus four divided by /two | equals \eight

    where, in a very simplified notation, | stands for IP boundary, / for a rising tone, and \ for a falling tone. Corresponding to:

    (6+4)/2 = 5
    6 + 4/2 = 8

  31. Amy Stoller said,

    February 7, 2010 @ 10:23 am

    @ Dan T: The hyphen lives on at the New-York Historical Society

  32. Stephen Jones said,

    February 7, 2010 @ 10:36 am

    you eat wild-mushroom soup

    Apart from Bill Walsh at the Washington Post, everybody eats wild mushroom soup. The mushroom soup may be really wild, but you're not going to find it in the wild, so the hyphen is unnecessary, just as it's unnecessary in zero tolerance approach. Where there's no ambiguity a hyphen is clutter.

  33. DonBoy said,

    February 7, 2010 @ 11:44 am

    John Hodgman has used "bases-ball" to similar archaic/humorous effect:

  34. Army1987 said,

    February 7, 2010 @ 1:23 pm

    Is "double beta decay" "(double beta) decay" or "double (beta decay)? Provided you take "beta" to be the name of the decay rather than of the emitted particle, both make sense and eventually mean the same. The latter makes slightly more sense to me, but the hyphenation Wikipedia uses in subsection titles implies the former. Also, the former would imply an Italian translation "decadimento doppio beta" and the latter "doppio decadimento beta", which give similar numbers of ghits (6160 vs 9050).

  35. mollymooly said,

    February 7, 2010 @ 2:35 pm

    "State Department Public Relations Director" might be tricky for what I believe are called "international" students, who are more likely to know that the USA consists of states than that there exists an entity called the State Department. Of course, only the most clued-in applicants are accepted at the OP's fine institution.

  36. GAC said,

    February 7, 2010 @ 3:11 pm

    @ marie-lucie

    I see similar problems with Spanish-language translations in the US. Pulling out a couple boxes from my closet for examples: I see "inkjet printer" translated as "impresora de inyeción de tinta" (though I understand that is the formally correct translation).

    There are some things that help: for example, using brand names as modifiers ("Original HP inks" became Spanish "las tintas originales HP", and French "l'encre HP origine"), and that Spanish at least can at least occasionally use plain nouns as modifiers, but sparingly (A cookware set labeled "Tri-ply Stainless Steel Cookware Set" could have had a slightly more aimless translation that "Juego de Cocina de Acero Inoxidable con Base Triple Capa [sic – capitalization]" if the author hadn't gone to the trouble of consolidating that final noun phrase a bit.)

    Of course, You can get similar effects translating other languages into English. Should 中华人民共和国 (China-populus/people-republic) be People's Republic of China, China People's Republic, or Chinese Popular Republic?

  37. marie-lucie said,

    February 7, 2010 @ 4:14 pm

    Stephen Jones:

    Where there's no ambiguity a hyphen is clutter.

    For literate, fast-reading adult native speakers who can process a whole sentence at a glance there may be no ambiguity, but there can often be ambiguity for beginning or slower readers, including foreign readers. Translators are usually in the second group, and amateur translators can't be expected to detect all instances of ambiguity, or to immediately resolve the ambiguity. Of course, spelling conventions are well-known to fully competent readers, but not every reader is fully competent. Hyphens do help.

    Mushroom soup: my example did not have just wild mushrooms (which would have been unambiguous) but the more complex wild forest mushrooms, so the point was whether forest went with wild or with mushrooms.

  38. Catanea said,

    February 7, 2010 @ 5:22 pm

    à Marie-Lucie – Ought you to declare a special interest in questions of hyphenation?

  39. Army1987 said,

    February 7, 2010 @ 5:57 pm

    Chinese People's Republic?

  40. Jerry Friedman said,

    February 7, 2010 @ 11:07 pm

    @Stephen Jones:

    Apart from Bill Walsh at the Washington Post, everybody eats wild mushroom soup.

    For some of us, that's only because we're not given a choice. (Not that I often get the chance to eat wild mushrooms, or especially prefer them when I do.)

    The mushroom soup may be really wild, but you're not going to find it in the wild, so the hyphen is unnecessary, just as it's unnecessary in zero tolerance approach. Where there's no ambiguity a hyphen is clutter.

    What do you mean by "clutter"? Does it interfere with your comprehension, the way clutter in a house (such as mine) interferes with one's ability to find things? Or is your objection aesthetic? Or something else?

    Anyway, there's a great deal of redundancy in punctuation. We could probably do without periods (full stops) as long as we keep capitalizing the first words of sentences. Faulkner, Thurber, and McCarthy have shown that we can do without apostrophes. But most of us include the punctuation anyway to avoid distracting the reader and to provide a little redundant help, since the cost is so small. If hyphens are different, maybe it's because we're used to doing without them.

    And as Marie-Lucie pointed out, not everyone is a fully proficient reader who doesn't need any help. Some children or students of English actually are seeing "zero-tolerance policy" for the first time and may find the hyphen useful (just as some people may find my gloss of "periods" useful even though I imagine you didn't need it). It's also easier for the writer to follow syntactic rules than to try to figure out whether there's any ambiguity possible in each case. And until now, I'd have said the hyphen costs practically nothing—but I hadn't thought about clutter.

    By the way, I'm not sure what "Wild Forest Mushroom Soup" is, but I suspect removing "Wild" would improve the name.

  41. codeman38 said,

    February 8, 2010 @ 9:34 pm

    On a similar note to Marie-Lucie's "mildew proof", I saw a package of batteries recently which translated "ozone free manufacturing" as "l'ozone libèrent la fabrication." Even worse, this appears to be taken directly from Babelfish, which translates it identically.

  42. Robert said,

    February 12, 2010 @ 11:10 am

    @Skullturf Q. Beavispants

    It's not just beginning algebra & analytic geometry students who make errors like that.

    Leibniz originally thought that d(xy) = (dx)(dy), before he realized his mistake and invented Leibniz's rule. Vladimir Arnol'd speculates in a book of his that Newton and Wallis couldn't make this mistake because they thought more geometrically.

    Also, Frobenius famously showed us that, at least in a ring of characteristic p, (x + y)^p = x^p + y^p.

  43. tk said,

    February 13, 2010 @ 9:06 pm

    My local Try 'N' Save has a special on "Australian Tenderloins."

    I asked if they were tenderloins of Australians.

    [Note, no species of meat was described in the label.]

  44. AlexTheSeal said,

    February 19, 2010 @ 2:51 pm

    Strongly agree with the pro-hyphen crowd here. I was taught that the hyphen binds the compound (adjective, noun, whatever) together on the left, and without it the phrase is not ambiguous but binds on the right. Thus a "stone traffic barrier" is necessarily a traffic barrier made of stone, but a "stone-traffic barrier" is a barrier to stone traffic (perhaps a means of preventing the theft and resale of rocks from geologically interesting sites such as the Grand Canyon).

    Around here (the Pacific Northwest USA) the yuppie supermarkets advertise "Free Range Turkeys" every Thanksgiving, meaning, I like to think, special turkeys which one can cook on the stovetop, leaving the oven free for yams and pumpkin pie.

    (A close paraphrase of Nicholson Baker, a writer who is nothing if not careful about semantics: "My first copy editor was a zealous anti-hyphenist, and in the proofs of my first book I had to write "stet hyphen" so many times in the margin that I finally began to abbreviate it 's.h.'")

  45. AlexTheSeal said,

    February 19, 2010 @ 3:52 pm

    "Meaning that they are giving away special turkeys," I should have said.

  46. Aaron Davies said,

    March 2, 2010 @ 9:42 pm

    the "new-york historical society", already mentioned above, is particularly nice, since it actually disambiguates two perfectly reasonable parsings: it's the (((new york) historical) society), not the ((new york)(historical society)). (incidentally, that style of adjective usage, much like "game-theoretic approach", seems somewhat (charmingly) archaic. i have a hard time imagining a linguist describing himself as being in the "computer-scientific department".) (also incidentally, didn't a linguist invent LISP?)

  47. Daniel said,

    March 23, 2010 @ 12:43 pm

    As just a regular ol' English major with little education in linguistics, I don't feel I can add much to the conversation. I actually came here via a search on Dogfish beer.

    What I can tell you is that the craft beer world is full of "parodied wine talk," and I expect it only to get worse. Can't we say nose instead of bouquet, at the very least?



  48. Peter G. Howland said,

    November 21, 2010 @ 4:47 am

    @Daniel – While I fully appreciate insider jargon as being a useful method of communication within any specialized industry, I have always been extremely put off by the term “nose” when referring to the smell of a particular wine or beer. When I hear the phrase “has an interesting nose” it suggests to me that what I’m about to drink contains an intriguing narene body part or its related snortage. I immediately think of a dangling poose of viscous mucoid exudate about to be honked into my glass because of the beverage’s odor. It’s like saying, “this pasta has a nice tooth” or “these shoes have a nice foot” or “these underwear have a nice testicle”. Blechhh!

    If you’d begin your reference to the assessment of the smell of a wine or beer with “aroma” you could then go on to describe its particular olfactory characteristics in whatever flowery terms are applicable. Such as: “In the aroma of this unique brew from Daniel’s Dungeon one can detect a latent bouquet of freshly regurgitated skunk vomit laced with not unpleasant wafts of finely pummeled patella from the crushed knees of malnourished urchins. Tastes pretty dang good, too!”, or whatever.

RSS feed for comments on this post