Job Discussion Forums

[fluffyhamster]

Came across this review whilst searching for Chinese frequency dictionaries on Amazon (and also found that a team including Tony McEnery are publishing just such a new dictionary later this year!):

[teacheringreece]

I'm planning on doing a lesson using corpora next week with an FCE level class. I'm thinking of doing something like a webquest (i.e. where Ss have a series of factual questions they have to find the answers to using the internet), but based on identifying structures, collocations, colligations, semantic prosodies and so on of various words. So I plan to produce a worksheet where they investigate the structures associated with a series of words and phrases, such as 'suggest' (here in Italy the common error is to use an infinitive with 'to' after it) and so on.

There's a very interesting website called Just the Word ( http://193.133.140.102/JustTheWord/ ), which is a free online corpus but seemingly specially tweaked towards language learners and teachers. It has a concept of 'bad combinations', so if you search for 'big shoulder' it will tell you that it's a bad combination and will suggest alternatives such as 'broad shoulder' and so on. I also like sketchengine, which is an incredibly powerful and quite simple way of researching a number of corpora, including the BNC. You have to pay, but you can sign up for a one month free trial.

The observation made earlier about 'green tree' not being a common collocation is an interesting one. I suppose similarly 'round circle' wouldn't come up very often in a corpus search (even though most of us would associate 'circle' and 'round'). The question is whether it is linguistically or pedagogically significant. Presumably we don't talk about green trees, or round circles, because there is simply no need, and for the same reason there's probably no need to worry about trying to teach students to avoid them as collocations. It also reminds me slightly of Hoey's chapter 'A World Beyond Collocation' in 'Teaching Collocation' (ed Michael Lewis). He seems to suggest there is significance in the fact that, for example, the word 'actress' doesn't occur in a corpus in a 'possessed construction', whereas 'architect' does; so we can say things like 'I'll speak to my architect about it', but one cannot usually 'possess' an actress in any similar sense. Hoey's point really is that we might at first expect all profession nouns to behave similarly in terms of colligation, but that this is not in fact true. My question though is whether we have to worry about it. If a real world 'fact' like the fact that most people generally don't employ actresses but many do employ architects has an effect on linguistic choice, do we have to worry about it? Or can we assume that being a NNS is not going to get in the way of this restriction imposed by the real world?

[johnslat]

"Presumably we don't talk about green trees, or round circles, because there is simply no need . . ."

True, but we might say, A circle is round" as part of explaining "roundness,"
or "The leaves are green" to illustrate "greenness."

As for using a "possessed construction" with architect but not actress, isn't it simply the case that there are certain professions that we can "hire",and/or that we have a personal connection with - my doctor, my teacher, my lawyer - and others that we can't and don't - my policeman, my actor, my pilot (well, unless you're really rich?)
As a side-note, do many people even use the word "actress" anymore? I thought, like most of the other "feminine -ess" words (with the exception of nobility: princess, duchess), it had fallen into disuse.
Regards,
John

[teacheringreece]

We can say 'a circle is round', but why would we ever need to? I suppose an EFL teacher might do that, but in terms of overall usage in a multi-million word corpus I doubt it's a significant collocation. 'Green leaves' is certainly more likely, but that's because we also talk about 'brown leaves', 'yellow leaves' and so on; I'm not sure the same applies to trees.

You're absolutely right about the hiring / personal contact aspect of the professions. But my point was that Hoey seems to imply that there may be a pedagogical significance in that, whereas I think as teachers we can expect certain real world facts to apply to NNSs as much as they do to NSs, so I don't think we have to worry about the fact that a corpus does not contain any examples with 'actress' in a possessed construction.

[fluffyhamster]

[Justin Trullinger]

One thing the corpus often does is points out the very real differences between real and display language.

"The circle is round," though an acceptable sentence, doesn't come up that much. Because "circle" means, um, round. But teachers are forever SAYING things to students, and teaching students to say things, that are about that unlikely.

"WHat is this?"

"It is a cat!"

"Is THIS a cat?"

"No, its a dog."

And other such drivel.


Whereas in real life, somebody who asks you if an animal, whether present or in a picture, is or is not a cat, is clearly, ahem, differently abled.

THough how "realistic" a corpus could be judged to be depends a lot on where the samples that make it up were taken from, it seems rare for display (meaning non-communicative demonstration of form) to find its way in.

Do NOT read CUshing's book. Important stuff, but I was NEVER afraid to fly until I started working in Aviation English. Nowadays...hard to handle. Lately, I drink too much when I fly.


Best,
Justin

[johnslat]

Dear Justin,
I was always afraid to fly after I became certified in air traffic control and ground control approach in the Marines.
Ignorance can be bliss.
Regards,
John

[fluffyhamster]

I wouldn't mind those sorts of silly questions if the picture (or its design) left some room for doubt*, and HEY, let's not forget about 'that' (i.e. where some degree of distance or trouble fully seeing might be involved). But you'll always get some teachers defending dodgy discourse by saying that students will be isolating and gaining useful phrasal stems/gap-fills such as 'Is this _____?' from dialogues such as the following (whilst overlooking the unlikelihood of 'Is this a _____?', unless one really IS meaning/wondering, "Is this common object (called, with an equally common noun) a ____ in English?", or on the other hand intending to supply more advanced phrases after 'Is this a _____?', such as 'good idea', or 'p*ss-take or what, this "Is this a goldfish?"?!'):

Sensei: (holding a picture of a dog) Is this a cat?
Me: No, that is not a cat.
Sensei: Is this a goldfish?
Me: No, that is not a goldfish either.
Sensei: What is it?
Me: It is a dog.

(The above is an English translation of part of a JFL lesson that really occured, apparently!).

*Something vaguely related:
http://forums.eslcafe.com/teacher/viewtopic.php?p=29583#29583

[Justin Trullinger]

The purpose, maybe, of the whole corpus thingy is to see how we really use language. The weakness is that it only really shows how we really use language in the material that got into the corpus.

It's been really useful in terms of showing what we teach, and model, that's actually infrequent.

But the other side is that if we get prescriptive with the corpus, we discount so many subsituations that never got transcribed.



To demonstrate, I just put the work "preposterous" into a well known corpus. It's a low frequency word, used only about one per 10 000 in the sample.

Then I put in the word $%&#. (Think of a very common offensive verb.) Much lower frequency. 1 in 100 000.

So according to this corpus, real English is to say "preposterous" five times more often than we say "$%&#."


$%&#ing preposterous, in my opinion.



Of course, to prove the point, I might have done this with a corpus with a mild bias towards academic writing. But you can see where I'm going. We have to be careful not to let the corpus become the next Azar.


Best,
Justin

[fluffyhamster]

A few threads that touch on CL:
http://forums.eslcafe.com/teacher/viewtopic.php?p=40194#40194
http://forums.eslcafe.com/teacher/viewtopic.php?t=8729
http://forums.eslcafe.com/teacher/viewtopic.php?t=8217
http://forums.eslcafe.com/teacher/viewtopic.php?t=7543
http://forums.eslcafe.com/teacher/viewtopic.php?t=7815
http://forums.eslcafe.com/teacher/viewtopic.php?t=5271

[fluffyhamster]

The lack of swearing is probably more indicative of an imbalance of purely (and generally pretty pure, clean!) written/printed sources over spoken, or even written-to-be-spoken, sources (e.g. Tarantino scripts), rather than any "getting prescriptive with the corpus" (that is, we tend as writers especially to do the "editing out" of inappropriate or excessive crassness ourselves; the corpus compiler doesn't then "need" to do anything, and certainly not "for" us); actually, I find the concept of 'corpus censorship' a bit of an oxymoron (do such views exist? If so, who's expounding them?!).

Fortunately, in ELT, despite the seeming overabundance still of the dodgier kinds of display language in bad teacher talk (which we should distinguish from e.g. off-the-cuff definitions that answer student queries), and there being quite a few lame textbooks around, modern corpus-informed textbooks have made some attempts to capture and relay the language more as it is; I'm not sure that the task-based methodology of for example the COBUILD course was entirely successful (commercially or even pedagogically), but it was good that they assembled a subcorpus of existing ELT textbooks, identified the difference between (i.e. what was lacking from) that sort of language compared to the wider "whole" (as contained in and revealed by the larger main corpora), and improved at least the selection of items to be taught if not the general pedagogy. (There is actually a link to a free pdf copy of the whole of Willis's The Lexical Syllabus in one of the threads there in my last post immediately above; see also the comments that the Willises made in Schmitt & McCarthy's Vocabulary: Description, Acquisition, and Pedagogy regarding the relative commercial failure of the COBUILD course (they were the authors) - click on the page 268 then 322 links here: http://books.google.co.uk/books?id=Epvtt9_zab4C&q=willis - as well as Sinclair & Renouf's 'A lexical syllabus for language learning' (Edit: Now available online here: http://www.wordtrends.org/publ/Lexical_Syllabus.pdf ) in R.Carter & M.McCarthy's Vocabulary and Language Teaching. Lastly, details of the design and exact composition of COBUILD's corpora, as well as of COBUILD's history and development generally, can be found in Looking Up: An Account of the COBUILD Project, a collection of interesting papers edited by COBUILD's founder, John Sinclair).

[MO39]

[fluffyhamster]

Some interesting pdfs available here:
http://www.corpus.bham.ac.uk/corpus-building.shtml

Note also the very valuable COBUILD Grammar Patterns 1: Verbs available in the 'Publications' section of the above site (I've mentioned this resource before though, in another thread).