Thursday, 14 January 2010

detecting errors

When proof-reading it can sometimes be an advantage to be a non-native speaker of the language in which the document in question is written. In our own language, we are often too dazzled by our familiar expectations to notice occasions on which they are not fulfilled. Faced, for example, with They want to to arrange payment we too readily normalize the text and perceive just the correct They want to arrange payment. But a NNS is more likely to be brought up short by the puzzling grammar of repeated to, and hence to detect the error.
Furthermore, it is typically NSs, not NNSs, who get confused about spelling problems such as there – their – they’re, your – you’re and its – it’s. As NSs we know what a word should sound like but may be uncertain how to represent it in writing. NNSs, on the other hand, are likely to be very aware of translation correspondences between the L2 and their native L1, where the English confusables are quite distinct.
I think I have already previously expressed my thanks to Masaki Taniguchi for repeatedly detecting various minor careless errors that I have committed in my blog postings over the years — typically, elementary typos or repeated words (often from when I have changed my mind about the best wording and end up with fragments of two different versions). As a NNS he can look at my effusions more objectively than I can myself.
These thoughts were prompted by a job I was doing yesterday. I was checking the draft of a document in Welsh, written by a Welsh native speaker. I am of course not a NS of Welsh, but learnt the language as an adult, by self-study and in evening classes in London.
The long and the short of it is that yesterday I was able to detect a fair number of errors of grammar and spelling that the author had overlooked in his own composition.
From my days in the classroom when I was learning Welsh I can remember one misguided teacher who chose to drill us learners on the distinction (in writing) between yw ‘is’ and i’w ‘to his’. This may be a tricky point for native speakers of southern Welsh, for whom the two expressions are homophonous. But it was not a problem for us NNSs: it would have never occurred to us to confuse them, since their translation equivalents are quite distinct. On the contrary, for some of us it might have come as a revelation to be told that the two are pronounced identically and could be confused by less-than-fully-literate NSs.
This is a special case of the more general principle that we can see other people’s mistakes more easily than our own. Get someone else to proof-read anything you write for publication. Ideally, perhaps, get two people to do it: one a NS and one a NNS.


  1. Of course, a speaker of North Welsh would not be likely to confuse yw [ɨw] and i'w [iw]. ;-]

  2. John,

    I bet you proof-read that post very carefully.

  3. I'm sure Hartmann-Skitt-McKean will have their revenge anyway.

    I've certainly noticed that the more fluent my English is (for some measure of fluency), the more of those substitutions I make. (I did indeed have to check "their" up there.)

    Similarly the more spelling pronunciations I rid myself off (/wraɪt/) the more errors I make where I would formerly have been aware of the the silent letter.

  4. The Welsh document in question was proofread by at least 5 Ns of Welsh; one a professional, university academic and another a licensed translator and they all (including myself, much to my embarrassment, as I am usually neurotic and anal about correct grammar and spelling) missed some errors which were glaringly obvious once pointed out. I can only claim in my defence that the errors were committed at 5am when my eyes were starting to glaze over.... and I am not accustomed to translating into Welsh - this was my first although I have several translations in print in English (from French) (ok... it was worth a try) However, John is quite correct as the most significant error corrections came from NNs, one of whom had a very rudimentary knowledge of Welsh. And even I was surprised at the confusion that (often competent and literate) Ns displayed in distinguishing between "a", "â", "ar", "â'r" and "a'r".

    John, I will thank you for your assistance once I have recovered from my neurotic breakdown ;-)

  5. A problem here in Worcestershire is when people confuse the spellings of "our" and "are". In the local accent, they sound just the same.

    I have yet to observe anyone in S. Wales mixing up the spellings of "year", "ear", and "here", which in S. Wales all sound the same as each other.

  6. As another "NNS", I have to support Sili's comment above. In at least some of these cases (i.e. those based on homophony, such as there/their above but not the repeated to), NNSs detect the errors because the pairs in question are NOT homophonous for them. I can attest that for a majority of my students, even very advanced ones, there and their simply do NOT sound the same. And once they do, they do so at a much more conscious level.

    (Said a commenter without much empirical support -- some nice research questions here -- anyone? Or has there been research already? I probably shouldn't admit I don't know. There certainly has been research into this kind of thing in "NSs", cross-dialectally, e.g. by Treiman et al.)

    And there are some sorts of these substitutions, such as sort-after for sought-after -- see comment by Kenny Easwaran on Language Log here and the Eggcorn Database here -- that no NNS in their right mind would ever make... That is to say, an NNS in an EFL context who first encounters most of their English in written sources.

    As always, things may be different for ESL contexts, especially with immersion. And for the pesky Scandinavians, Sili, who get their TV in subtitled English rather than dubbed into their L1 :)

  7. The a/ac ("and") vs â/ag ("with") confusion seems to occur most often in one construction, namely the Ymadrodd Annibynnol ("Absolute Clause"?), which in Welsh uses a/ac, but in English is usually translated with "with". Google indicates a 3:2 ratio between "ac yntau'n" and the non-standard "ag yntau'n". Although this error does show up elsewhere (e.g. "aethpwyd ac ef" provides a handful of results), genuine confusion seems to be pretty restricted.

    And I'm completely mystified as to how anyone could ever confuse "yw" and "i'w", although it might begin to explain why language teachers tend to like "ydy" so much. I'd've thought that the more common error would be turning "i'w" into "i" (which often follows the English pattern) or "ei", as I'm not sure I'd actually pronounce the "w" in common phrases such as "rhywbeth i'w wneud".

    Actually, there's another confusion that does happen a lot. The words "i", "ei", and "eu" are all pronounced "i" (at least down here) -- "ei" is a spelling introduced by William Salesbury in the Bible, because he thought it was the same word as the Latin "eius". I've lost count of the number of times I've seen first-language speakers muddle these up. One particularly virulent one is "eu gilydd" (and "'u gilydd") in place of "ei gilydd" (and "'i gilydd").

  8. David Marjanović16 January 2010 at 23:24

    A problem here in Worcestershire is when people confuse the spellings of "our" and "are". In the local accent, they sound just the same.

    I see Americans doing this every week.

    Google indicates a 3:2 ratio between "ac yntau'n" and the non-standard "ag yntau'n".

    Could it be that the |k| actually ends up voiced between the vowels in that phrase?