Bain on restrictive relatives

Over at Language Log, they've taken to calling the restrictive that rule "Fowler's Rule". And from what I recall, Merriam Webster's Usage also attributes this to Fowler (1926).

I was trolling Google books for more on coordinating conjunctions, when I came across Alexander Bain's An English Grammar (published in 1863 by Longman, Green, Longman, Roberts & Green, one of the various incarnations of Longman). A section in the introduction caught my attention (click for larger images):

This is followed up on p. 23 with the following:

Interestingly, in the examples that Bain supplies to show how which can be ambiguous, he seems to completely overlook the comma, which--at least from this twenty-first-century viewpoint--removes any trace of doubt at all.

So, even if Fowler popularised the rule, Bain claims it as a novelty over 60 years earlier.

The Same in Japanese

Kieran McGovern edits a web site with reading material designed for English language learners. He was recently asking around for materials, so I sent him a short story that I wrote with a language-learner audience in mind. The story, called The same in Japanese is here. If you use it with your students, please let me know what the reaction is.

With 'should', expect the best

Listen to this article

Someone over on the ETJ list asks how to explain the problem with should in the following sentence: If you fly on a Bombardier airplane, you should die. The intent is not to express some kind of moral outrage about Bombardier, but to suggest that their airplanes are likely to drop from the sky at any moment. Yet, that's not how most native speakers of English would interpret the above.

I'm not really sure why this should be so, though I believe that it must depend to a certain extent on the ability of the addressee to comply with the utterance. Compare:

• If you fly on a Bombardier airplane, you should arrive on time for your flight. (advice)
• If you fly on a Bombardier airplane, you should arrive on time at your destination. (prediction)

Given this, "...you should *expect* to die" fixes the problem, since expectations are, as long as you're pre-warned, somewhat within your control.

Where the verb denotes something outside of your volition, there seems to be a difference between positive outcomes and negatives (at least in this particular sentence.)

• A. If you fly on a Bombardier airplane, you should survive. (prediction)
• B. If you fly on a Bombardier airplane, you should die. (?)
• A. If you fly on a Bombardier airplane, you should be satisfied with the service. (prediction/advice)
• B. If you fly on a Bombardier airplane, you should be dissatisfied with the service. (?)
• A. If you fly on a Bombardier airplane, you should arrive on time at your destination. (prediction)
• B. If you fly on a Bombardier airplane, you should arrive late at your destination. (?)
• A. If you fly on a Bombardier airplane, you should have a pleasant trip. (prediction)
• B. If you fly on a Bombardier airplane, you should have an unpleasant trip. (?)

I'm not sure how far this observation can be generalised, but there it is.

Learning languages and lobes

Listen to this article

Merton Bland regularly posts to the TESL-L list. Unfortunately, little of what he has to say is helpful. One of his favorite arguments goes something like this: "If our job is to bifurcate the native and target languages, any instrument (such as a bilingual dictionary) which forges chains to tie the two together is counterproductive." He backs this up with arguments like this: "A true bilingual does his processing of each language in a different hemisphere of the brain." Sometimes he changes hemispheres to lobes.

If there's a need for "bifurcation", it is in separating evidence from ideology.

The claim that a second language is processed in the other hemisphere of the brain is completely absurd as the studies cited below will show. The idea that it is processed in a distinct lobe may or may not be more correct, but it remains misleading.

It is generally agreed that there are five major lobes in the brain. The definition of 'lobe', however, is rather ambiguous and minor lobes are discussed, but not always agreed upon. Furthermore, two areas of the brain that are mainly involved in language are Wernicke's area, which spans the region between temporal and parietal lobes, and Broca's area, which is in the frontal lobe. In other words, language activity usually occurs in different lobes in monolinguals, too, not just in bilinguals.

While late L2 learners do seem to activate different (though typically adjacent) areas of the brain, the research on the overlap or lack thereof of L1 and L2 brain activation is equivocal. A summary by Vivian Cook of some available studies is here.

Bland's point is that teachers should avoid connections between the L1 and L2 (through translation) to dissociate them in the brain. Yet, one idea that seems to come from these brain studies is that the most successful bilinguals are actually those who seem to have the greatest overlap, not the greatest separation. This might mean that connections are good, not bad. Or not. Either way, they are likely inevitable.

Personally, I don't think the brain studies are ready to tell us much about how we should go about teaching language, but they're fun to look at.

If Murdock was right...

Listen to this article

Back in February, Deroy Murdock wrote an article in which he quotes and then incorrects John Kerry.

"If I was president, this wouldn't have happened," John Kerry said during Hezbollah's war on Israel last summer. As 2004's Democratic presidential nominee should know, he should have said, "If I were president ..."

I wrote to ask Murdock what he bases his claim on, but I'm shocked to say, didn't hear back from him. So, I'll explain here why I think Kerry is right both grammatically and politically.

Murdock asks, "is it pedantic to expect linguistic excellence?" "No!" he says, and I agree, no. It does, however, seem pedantic to insist that were in counterfactuals is the excellent option.

We have been using was here for over 400 years, among the earliest-known instances being in the writings of Christopher Marlowe. It was common by the end of the 17th century and was thought acceptable enough to be employed by Defoe, Swift, and Addison and many other great writers.

A search of the British National Corpus (sorry, the American National Corpus isn't completed yet) shows if I was to be about 35% more frequent than if I were, though I have no automatic way of identifying which are actually counterfactuals. Still, it seems safe to say that counterfactual was is common both in speech and in edited prose.

Even prescriptivist usage guides support it. H.W. Fowler wrote, "Subjunctives met with today, outside the few truly living uses, are either deliberate revivals by poets for legitimate enough archaic effect, or antiquated survivals as in pretentious journalism, inflecting their context with dullness, or new arrivals possible only in an age to which the grammar of the subjunctive is not natural but artificial."

In fact, Kerry may have carefully chosen was over were. Where the present tense is the natural choice for clauses we view as facts, the past tense, when it is not indicating past time, often presents hypotheticals. Indeed, since be is the only modern English verb with a distinct irrealis form, most verbs actually rely on the past tense to indicate that the predicate does not currently obtain. But be, with its multiple possible forms, presents us with options. These options may simply be the choice between a formal were and a more colloquial was. But it may be more insidious than that. Formality is distancing, and such distance may create a perception not only that Kerry isn't president, but that he has no hope in hell of ever actually taking office. Possibly true, but not a proposition Kerry would be keen to voice.

FANBOWA?

My very first post on this blog was about FANBOYS. In that post I mused about where this mnemonic might have originated. This evening, as I was reading The Sound on the Page, by Ben Yagoda, I thought I had found my answer: Oliver Strunk. When Yagoda quoted from The Elements of Style, “place a comma before a conjunction introducing an independent clause,” I thought, "of course!"

Strunk illustrates with two examples—one with and and one with but—before continuing with the following prescript.

"Two-part sentences of which the second member is introduced by as (in the sense of because), for, or, nor, and while (in the sense of and at the same time) likewise require a comma before the conjunction."

In other words, Strunk, while largely agreeing with the spirit of FANBOYS, differs in the details. Firstly, where we currently have FANBOYS, he had FANBOWA. (For Strunk, so was an adverb requiring a semicolon; yet is not mentioned.)

Secondly, nowhere does Strunk write, "coordinating conjunction". Instead he distinguishes between conjunctions that join independent clauses, and other conjunctions that “likewise require a comma”. This suggests that he thought for, nor, or, while, and as introduced a dependent clause, though I can’t think why or and nor are listed here.

So The Elements of Style (Geoff Pullum has called it, a "vile little assemblage of stupid advice about usage") is a little confused about or and nor, but doesn't seem to be the source I was looking for. This clue does suggest, however, that FANBOYS has its origin sometimes after 1935, when TEOS came out.

BTW, Schoolhouse Rock had it right: and, but, or & nor.

Discriminatory etymology

Listen to this article

John Steckley, a professor of anthropology here at Humber College, has just released a book entitled, Full Circle: Canada's First Nations, from which he was kind enough to send me the chapter on Aboriginal languages. In the chapter he points out how ridiculous it would be to simply specify that a word is of European background. Yet that, he claims, is essentially what many dictionaries do when they give the etymology of, for example, moccasin as N. Amer. Ind.

My first reaction was "how true!" followed very closely by "really?" All of the dictionaries that I could find in the house and online gave reasonable etymologies for moccasin. For example, Dictionary.com gives, "Origin: 1605-15, Americanism; - Virginia Algonquian - Proto-Algonquian *maxkeseni." Steckley gives sources, but since I only had chapter 4, I didn't have the bibliography and couldn't follow them up. I was suspicious.

This afternoon, we were down at the bookstore with the kids, so I decided to check out some other dictionaries. Sure enough, two or three had etymologies as general as "Amerind". Most Oxford dictionaries were fine, but the Oxford School Dictionary, marketed to children, had the trivialising etymology--this, despite having Greek and Dutch elsewhere. Surprisingly, it wasn't just dictionaries for the junior linguist that were at fault. The Collins Canadian Paperback Dictionary was fine, but the Collins Canadian Dictionary and Thesaurus, a big, seriously adult-looking hardcover wasn't.

Steckley does a good job clearing up the ridiculous notion, inexplicably propagated by these dictionaries, that there is but a single American Indian language, and then goes on to take apart a number of other common misconceptions about the languages of Canada's First Nations.

Writing systems and learning to read

Listen to this article

Over at Language Log, Bill Poser oversimplifies things somewhat when he writes,

If you have a straightforward phonological writing system and teach children to make use of that structure, most children will learn to read and write without great difficulty. The idea that learning to read and write is a lengthy and painful process is a pathology of our writing system and educational system, not a universal truth.

There are certainly grains of truth in there, but in the case of English, the phonology of the language is largely to blame for our less-than-transparent writing system. Bill puts forward the eminently sensible "frog's feet" syllabic writing system for Carrier. Would such a system work for English?

Frog's feet has roughly 200 characters representing the intersection of 6 vowel sounds and 33 consonants (not 6 x 33 because not all can be used together). By contrast, if we take the counts from the Longman Pronunciation Dictionary by John C. Wells for General American English, we have 25 consonants and 19 vowels, with one additional consonant and three additional vowels for foreign words. With a few exceptions, all of these can be used in isolation (i.e., word-initially for vowels or, for consonants, before another consonant or as the final letter in a word.) Furthermore, most of the consonants can be followed by any of the vowels. (See here for more phonotactic rules.) I can't be bothered to count the exceptions, so I'll estimate by reducing the counts by 20%. Under this estimate, you would need somewhere between 300 and 350 characters to replicate the frog's feet system for English.

Moreover, the elegant simplicity of the Carrier system is impractical for English. My understanding is that frog's feet uses a basic shape for each consonant, and simply rotates these to indicate the relevant vowel (diatrics are also emloyed). When you have only 6 vowels, this works beautifully. When you have 19...

Finally, the Carrier system, with its occasional isolated consonant is mostly syllabic, but with alphabetic elements. And it is well documented that syllables are something that humans can fairly easily notice and learn to read. On the other hand, the English system, with its higglety-pigglety consonant doubles, triplets, and even quadruples (e.g., sixths), would remain largely alphabetic. Alphabetic systems require phonemic awareness, and this is harder to attain than syllabic awareness.

So, Bill's right about frog's feet being easy to learn, but to suggest that a similar system would actually be practical for English, let alone make it any easier to read, strikes me as unlikely.

[Japanese also employs a syllabic writing system (actually two parallel systems) along with the kanji logographs. The Japanese system is even more compact than the Carrier one as Japanese has a smaller complement of phonemes and all syllables are either V or CV except for tsu and n. Consequently, they have only 52 characters (along with one major and one minor diatric mark).]

Different kinds of questions

Recently I was speaking with a colleague about her plans for her new low-level class. When she said that she intended to focus on question formation for the first few weeks, I asked her with some trepidation how she would be sequencing the various kinds of questions. It turned out that she didn't know what the various kinds of questions were. And she wasn't the first.

Broadly, there are yes/no questions (e.g., Is it hot in here?) and so-called wh-questions (e.g., Where do you live?). Leaving aside minor forms such as tags (e.g., It's hot, isn't it) and echoes ( e.g., You're what?) Yes/no questions can be divided into those with a need for do support and those without.

If the corresponding affirmative form of the question has only a lexical verb and no auxiliary verb, then the question form requires the insertion of (do) before the subject.

• He likes this. -> Does he like this?
• He went to Halifax. -> Did he go to Halifax?

In contrast, if the sentence has no lexical verb or includes an auxiliary, then do support is not required. Instead, the non-lexical verb is fronted (i.e., moved in front of the subject).

• You will marry me. -> Will you marry me?
• He is nice. -> Is he nice?

A similar difference exists in wh- questions; lexical verbs cannot be fronted, so, where fronting is required, do or another auxiliary is needed.

• He plays in Etobicoke. -> Where does he play?
• He has played in Etobicoke. -> Where has he played?

But, not all wh- questions are the same. The wh- words come in three flavours: pronoun (who, what), determinative (which, what), and adverb (where, when, why, how). The adverbs always require an auxiliary verb before the subject.

• Where did he go?
• How should I do it?
• Why is that happening?

In contrast, question formation with the pronouns and determiners differs depending on whether the question is about the subject or about some other noun within the sentence. If the question relates to the subject, the result is the canonical subject-verb-object sentence.

• The door went bang. -> What went bang?
• Brett can get the pizza. -> Who can get the pizza?
• This pencil is the best. -> Which pencil is the best?

Otherwise, you end up with the same situations as above with the adverbs.

• I banged the door. -> What did you bang?
• Brett can get the pizza for us. ->Who can Brett get the pizza for?
• She likes that pencil. -> Which pencil does she like?

With all this to take into account, I wonder how a teacher could possibly plan a coherent set of lessons on question formation without understanding how questions are formed. Make no mistake; I don't blame my colleague. But I do worry about TESL training programs.

Teacher training big on how, missing the what

Listen to this article
A few weeks ago Bill Poser claimed that (Canadian and US) teachers aren't taught how to teach reading.

"Even in states and provinces where phonics is used, teachers are generally not given the training necessary to teach it. Here in British Columbia, for example, students training to be teachers read a little bit about approaches to the teaching of reading and discuss them, but they are never taught what the sounds of English are, what the letter-to-sound rules of English are, or how to bring about phonemic awareness. Teachers who lack this knowledge are hard put to do a good job of teaching phonics."

I asked a colleague who had recently completed teacher's college here in Ontario whether this reflected her own experience there; she basically concurred with Poser.

"In teacher's college we learn how to teach different reading strategies, but we don't learn a lot about phonetics or how to teach (it). I guess in 8 months with placement included there is not enough time to go in depth into everything."

The other day, the Toronto Star had an article indicating that there is a glut of applicants to teacher's college and of recent graduates. This will likely have some positive effect on the quality of the individuals coming out, but without the education and training, there's little hope that these folks will be any better equipped to teach reading.

(By the way, I remember hearing of some irreverent wag, commenting on the difference between training and education with a simple question. "How many of you would accept sex training in public schools?")

Grammar course for ESL teachers

Listen to this article
The Guardian is touting an online grammar course for prospective ESL teachers offered by a private training company called Cactus. There's no doubt that more teachers need such a course; I just hope that they don't end up learning that there are no verbs in phrases and other similar nonsense. The course is accredited through the Open and Distance Learning Quality Council, an organisation that I'm unfamiliar with, but which appears to be legitimate.

John Sinclair (1933-2007)

Listen to this article

Elena Tognini Bonelli wrote the following obituary:

Yesterday was a very sad day for the world of linguistics. John Sinclair (b. 14 June 1933) died at his home in Florence, aged 73. He will be deeply missed by his family, his colleagues and his many friends. His death is a terrible loss to everyone who knew him. Friends and colleagues are welcome to the interment of the urn which will be at the Cimitero degli Inglesi in Florence on 28 March at 3.30pm.

John was an outstanding scholar, a first-generation modern corpus linguist and clearly one of the most open-minded and original thinkers in the field. He was Professor of Modern English Language at the University of Birmingham for most of his career and founder of the ground-breaking COBUILD project in lexical computing which revolutionised lexicography in the 1980s and resulted in a new generation of corpus-driven dictionaries and reference materials for English language learners. After his retirement from Birmingham John moved to Italy where he became President of the Tuscan Word Centre, an association devoted to promoting the scientific study of language. On the short intensive courses that the Tuscan Word Centre offered, John very generously shared his original ideas about language and linguistics with generations of younger scholars, introduced numerous students to the fascinating world of corpora and inspired many new ideas for future research in linguistics. He was an Honorary Life Member of the Linguistics Association of Great Britain and a member of the Academia Europaea. John held an Honorary Doctorate in Philosophy from the University of Gothenburg, and Honorary Professorships in the Universities of Jiao Tong, Shangai and Glasgow.

He is gone now and it will be very hard to get used to it. John's last email to me just a couple of days ago ended "Very brief note tonight; more to follow." I will miss him.

DST

John Sakamoto, citing webexhibits.org, writes,

"Daylight Saving Time is grammatically correct. "Saving" acts as a participle, which modifies "time" and tells us about its nature. However, the phrase itself is inaccurate, since no daylight is actually saved. It is merely shifted from morning to evening."

Daylight Saving Time is indeed grammatically correct, as is Daylight Savings Time. The analysis is, however, faulty as the caveat implies. If anything at all is being saved, it's daylight, not time. So, to say that 'saving' is modifying 'time' strikes me as very odd.

I see two possible analyses. One is that it is time for saving daylight, analogous to time for reading the paper. Under this strategy you get daylight-saving time, analogous to paper-reading time. The s doesn't work here.

The other strategy is to analyse saving(s) as a noun, as in a one-time saving or a savings account. Under this strategy, daylight is a noun modifying the second noun saving(s). Now, if we look at the analogous cost saving(s), we find that the plural is more than twice as common as singular, but neither is rare.

But we're not finished. The noun phrase, daylight saving(s) is itself modifying time. While there are cases where a plural noun is used as a modifier (e.g., antiques show), it is far more common for such modifiers to be singular. Indeed, with our analogous cost saving(s), we find cost saving measures is roughly five times as common as cost savingS measures. Still, this doesn't rule out the possibility of the plural form if one wants to emphasise that there is a saving every day, as discussed here.

By the way, the daylight is saved for many individuals who would otherwise be sleeping through it in the morning, thus wasting it.

To revivify or condemn

Listen to this article

Over at Language Log, folks are wondering aprint (here and here) how prescriptivist rules get formulated in the first place. What moves a usage raven maven to take umbrage at something that has seemed perfectly innocuous to generations of English speakers?

Well, the other day, I came across revivify in a long but fascinating NYT Magazine article. I had never seen the word before and read it \ri-ˈvīvə-ˌfī\ instead of the correct \rē-ˈvi-və-ˌfī\. My reflex reaction was "what a silly and pointless neologism! What's wrong with revive?". But a quick search of the dictionary, reveals it's nothing of the sort.

While revive is older (circa 1400), revivify is attested from 1665. Moreover, English speakers have been able to vivify something (another new word for me) since 1535. That's a trick that revive can't do. And while both revive and revivify mean bring back to life, revivify seems to have a link to vivid that revive is simply lacking. Still, revivify is fairly rare, appearing almost never in speech and only about once every 10 million words in print.

So, here I am at a crossroads. Do I abashedly admit that I was ignorant and unobservant and welcome revivify into my vocabulary, or do I try to save face by condemning it as a mere florid redundancy? The second choice, pathetic though it is, may be one possible avenue by which language items fall into disrepute.

A woman the source of the 'he' rule?

In following up Arnold Zwicky's by-the-way attribution of the rule that only he should be used as an indefinite pronoun (e.g. Every person should make sure he...) to Lindley Murray in 1795, I came across an article, which antedates that instance by 50 years and attributes the rule to Ann Fisher.

Synesis across the Atlantic

I promise this is my last post on Richard Firsten's Grammatically Speaking column for quite some time. Again, in his most recent column, he takes a question about grammatically singular nouns, such as team and crew, with plain present tense verbs instead of the third-person s forms. This is called synesis (discussed earlier here). In response, Firsten writes

In a way, the BrE is more consistent than AmE. AmE has the police are and people are, and it's not so unusual to hear the singular or plural verb form after family. When talking about the family as a group of individuals, AmE tends to use the plural verb form; when thinking of the family as a whole unit, AmE uses the singular verb form.

I'm afraid, I don't see Richard's point about police and people. Perhaps he is confusing irregular plurals with singular collective nouns? Either way, it's not just AmE; You would never find the police is or the people is in BrE either, where police and people are the subject.

And is family as group versus a number of individuals really an AmE issue? In the British National Corpus, the string family + (3rd person singular verb) is over 1.5 times more common than the string family + (plain form verb). Not all of these are relevant, of course. They include sentences in which family is not the subject (e.g., The problems with the family are numerous & The nice thing about this family is...). Still, this distinction is clearly available to speakers of English on both sides of the Atlantic and there's no evidence in Firsten's column that BrE is more consistent than AmE (though it may be).

Another of Firsten's fumbles

Forgive me if I appear to be picking on Richard Firsten. It's just that if the guy who answers teachers' grammar questions for the largest organisation of English language teachers in the world can't do better than here, here, here, and here, then we're in trouble.

In his latest column, he agrees with Gail Sepúlveda of Monterrey, Mexico that the following are partitive genitives:

a pair of pants
a cup of coffee

Really? A partitive construction is one in which there is a part to whole relationship, as in some of the people. In this example, there is a set of people and some identifies a subset of those people (i.e., one part of the group). The same thing applies to non-count nouns, for example a little of the wine.

But that's not the relationship in either of the above phrases. Consider the difference between a pair of the pants and a pair of pants. The first is partitive, the second non-partitive. The same applies to the cup of coffee example.

To be honest, I'd never paid any attention to the term partitive though I know I've seen it a number of times before. Yesterday, I wouldn't have been able to tell one from a flux capacitor. But it just takes a few minutes' effort to find out. Unfortunately, too many of us English language teachers seem to lack the curiosity or wherewithal to identify and use good references.

Headed back again

Back in December, I discussed Richard Firsten's analysis of "She's headed for the beach." In his newest Grammatically Speaking column, he tries to explain that in the last column, he didn't mean that headed was an adjective, but merely that it was being used "adjectivally". It's not really clear what he means here, but it seems that, following this example, he would consider all participles in passive constructions like 'he was eaten' and all in progressives like 'he is eating' to be an "adjectival phrase". I don't find this either convincing or helpful.

Richard goes on to say that "is headed for the beach" is a statal passive. According to Huddleston's Introduction to the Grammar of English, "a statal passive ... attributes a certain property to the (subject), that of being in the state resulting from (an) event." Huddleston goes on to say that the participles in statal passives can also appear after other copulative verbs, as in "the vase appeared / seemed / looked broken." Neither of these properties hold true for our headed example. Being headed for the beach is not the result of an event, nor can we say "She seems headed for the beach" (a web search will turn up thousands of hits for "seems headed", but all appear to be headlinese.)

[Update: Upon further consideration, I think this IS a statal passive. I was having serious trouble dissociating the movement entailed in heading for something from the grammatical properties of the sentence. After some thought and research I now think that a defining characteristic of statal vs. dynamic verbs is that you cannot refer to a point inside dynamic verbs without using the progressive. States, in contrast, allow such reference. So we have

• At 2:00 she was headed for the beach. (2:00 is a point inside the state)
• At 2:00 she walked to the beach. (2:00 cannot be a point inside the dynamic occurrence)

To my chagrin, I now believe my initial analysis was incorrect.]

How I learned English Grammar

One of the members of the ETJ list wrote, "I was wondering for those of you who teach grammar...how did you learn grammar?" I thought I might share my journey, such as it has been.

I learned a good deal of grammar in my French immersion school in the early 70s. Of course, that was in French. When I went to Japan and began teaching English, I had a number of basic concepts, but was often at a loss to understand and explain many points. I picked up some more grammar by staying one step ahead of the students in the various textbooks like that one they (used to?) use at Nova. Still, these were not really grammar textbooks per se and what I picked up there was limited.

My first two real grammar books for teachers were Michael Lewis's The English Verb and A Teacher's Grammar: An Approach to the Central Problems of English by R. A. Close, which I just stumbled across during my regular browsing at Kinokuniya.

Later I took Temple University's M.Ed. which includes a course called New Grammars taught by Ken Schaefer. Ken took us through more traditional grammatical analysis, including sentence diagramming, and introduced us to Chomskyan transformational grammar with grammar trees and all that. The main text was The Grammar Book, which is still probably the best course in grammar for English language teachers.

Since then, I have added to my collection the Longman Grammar of Spoken and Written English, which has lots of interesting stats, but suffers from a lack of attention to its theoretical foundation; and the Cambridge Grammar of the English Language, which seems to me the current pinnacle of analysis. (This is not to be confused with the terrible Cambridge Grammar of English by Ron Carter and Michael McCarthy.) There is also a smaller textbook for undergrads, A Student's Introduction to English Grammar based on the CGEL.

Finally, whenever there is a grammar question, I do my best to track down an answer and explain it. You learn a great deal teaching others.

Recursion discovered at New Scientist

New Scientist magazine generally does a very good job of reporting science news. They also have a section called Feedback in which people tend to send in random amusing stories, many of which have to do with language. Take, Olaf Lipinski's story about his new shirt:

The other day he bought a shirt with black and white stripes. He was delighted with it until he read the label which says: "Wash dark colours separately".

These are all well and good, but back on Jan. 20, there was the following:

Understanding understanding

READER Stan Courtney tell us that while browsing through Scientific American he came across this in an article about autism: "For neuroscientists, this finding... represents a dramatic change in the way we understand the way we understand."

Courtney says he cannot remember seeing any other sentence like this, in which a repeated phrase actually makes sense. Has anyone else?

The follow up appeared in the Feb. 10 edition:

The way we understand the way we understand

WRITING in response to Stan Courtney's comment on repeated phrases that make sense, such as "the way we understand the way we understand" (20 January), Stewart Haywood tells us: "I watched my daughter standing in front of a mirror this morning, looking at herself looking at herself."

...David Squire's simple submission "This sentence contains 'this sentence contains'" comes as something of a relief.

Many other readers offered their own versions of this linguistic oddity, but space prevents us publishing them all, so let us conclude with the wisdom of the physicist Richard Feynman, who was there before us in this, as he was in so many things. Tim Cowell was the first of many to draw our attention to a statement made by Feynman when he was a student, which went something like this: "I wonder why. I wonder why. I wonder why I wonder. I wonder why I wonder why I wonder why I wonder!" He repeated this to himself over and over again, apparently, as a way of sending himself off to sleep.

All of which makes you wonder: is this a phenomenon that has just caught the attention of one of the world's leading science news magazines? Do they really know so little about language and linguistics that they are unaware how fundamental the idea of recursion is? They have, after all, written about it before and here. Admittedly, this is a special limited type of recursion, but it's still pretty obvious, isn't it? I mean, my five-year-old son once told me a "story about a story about a story about a story..." and found it terrifically funny.

Can you imagine somebody writing in and saying, "I've recently noticed that light behaves both as a wave and as a particle"? But then I guess we expect people to know more about physics than about linguistics, a fact that Mark Liberman is always lamenting.

More here
A particularly disgusting (I'm not kidding!) example of this type of recursion is here (click at your own risk), other related Language Log posts being here and here. Escher made much use of recursion in his drawings, including the special case of the strange loop as illustrated in drawing hands. Finally, GNU is a recursive acronym which means GNU is not UNIX.

Google Scholar & free papers

Last week I was looking for information about using timed lexical decision tests to assess vocabulary. I forget exactly which search terms I used, but one of the papers google scholar recommended was Kempe, V., & MacWhinney, B. (1996). The crosslinguistic assessment
of foreign language vocabulary learning. Applied Psycholinguistics, 17, 149-183. For this paper, Google returned the following options: Cited by 10 - Related Articles - Cached - Web Search - BL Direct

None of these will actually take you to a copy of the paper unless you or your insitution are subcribed to BL Direct, which might lead you to suspect that there is no free version available. If, however, you go to Brian MacWhinney's web site, you'll find a free copy there for your perusal.

Still, I don't think there's any kind of sweetheart deal going on here. Danny Sullivan on searchenginewatch.com writes,

"Google says, by the way, that it does not earn money off of any new subscriptions generated between searchers and publishers."

And, in fact, if you do your google searches unflavoured, you can find the free version of the Kempe & MacWhinney paper easily. Its not even that Google Scholar never links to free papers, as the following results demonstrate:

Attention Control and Ability Level in a Complex Cognitive Skill: Attention Shifting and Second- … - group of 4 »
N Segalowitz, S Frenkiel-Fishman - Memory & Cognition, 2005 - www-psychology.concordia.ca
One of the most interesting research results of the last half-century regarding
skilled performance is that what primarily distinguishes experts from ...
Cited by 3 - Related Articles - View as HTML - Web Search - BL Direct

So there's something odd happening here.

Some more evidence of the weirdness is that GS doesn't seem to know everything that just plain google knows. For example, if I copy the first line of the abstract and search for that in quotes, GS just shrugs its shoulders, while bare bones google serves up the required paper with its usual alacrity.

What's up? Who knows, but if you don't find a free version of a paper on GS, don't give up. It may still be out there.