May 04, 2007

Back to the future, redundant preposition department

Yesterday I posted about an Australian student whose political manifesto began with a flourish of distinctly non-standard prepositions: "I am one of those Liberals with which this publication has a somewhat unhealthy obsession towards. This article would like to explore some issues to which this newspaper often propagates on." ("A note of dignity or austerity", 5/3/2007).

The problem here is not that that "towards" and "on" are placed at the ends of their clauses, but that additional prepositions are also provided with the clause-initial relative pronouns, "with which" and "to which". (Well, there are other stylistic problems in these sentences as well -- it's more usual to have an obsession with or about something than an obsession towards it; and you can propagate (i.e. spread) a religion, but it's idiosyncratic and distracting to write about propagating on some issues, as if propagate meant "preach" -- but this is not a critical-writing seminar.)

I agreed with Renae O'Hanlon's diagnosis, which was that the author had introduced "with which" and "to which" because they seem "proper and sophisticated, and give him an air of authority". And then I jocularly suggested that the redundant propositions, if adopted by generations to come, might eventually lead to the re-introduction of case marking on relative pronouns in English, through forms like "twitch" and "fwitch". The idea behind this joke was that inflections sometimes develop from re-analysis of phonologically-incorporated function words (see e.g. Arnold Zwicky and Geoffry Pullum, "Cliiticization vs. Inflection: English n't", Language 59 (3) 1983), and that redundant prepositions might be the leading edge of case-marking agreement.

I should have known that there's rarely anything new under the sun, at least when it comes to English morpho-syntax. It doesn't matter whether you view an "innovation" as a disgusting display of terminal degeneracy, or as an amusing example of linguistic creativity -- the chances are, Chaucer or his contemporaries did it too. This morning, I got a note from David Denison, reminding me of this principle.

David's note:

My DPhil dissertation, written back in the Lower Pleistocene, was about the history of phrasal and prepositional verbs in English (and Norse, re phrasal verbs), and at that time I collected quite a few examples from Old, Middle and early Modern English like the ones you cited. I doubt it's a change in progress, therefore, at least as far as the FORM is concerned. The speculation about _with which_ and the like being a new status marker is fun, though, so a new FUNCTION for the form is possible. Exaptation, maybe?

Anyway, a little Middle English evidence:

& sei me hwer þu wunest meast; of hwet cun þu art ikumen of (Seinte Marherete 38.1)
and tell me where you dwell most, of what race you are come of [of = 'from']

Till all & syndry to quham þe knawlage of þir present lettris sail to cum (quot. 1428, OED s.v. whom pron. 7a)
Also they found there namys of ech lady, and of what bloode they were com off (Malory, Works 593.24) -
[NB. Malory does it a lot, and there are more examples in Middle English Dictionary s.v. _of_ adv. 7.]

With different particle:

þuruh hwat muhte sonre ful luue of aquikien (Ancrene Riwle (Nero) 25.16)
through what might sooner foul love from awaken

Ne nis na þing hwerþurh monnes muchele madschipe wreððeð him wið mare þen þet schafte of mon (Seinte Katerine 234)
NEG not-is no thing where-through man's great madness angers him with more than that that creation of man

Unto which place every thyng, . . moveth for to come to (Chaucer, Hous of Fame 733)

And Nuria Yáñez-Bouza, one of my present PhD students working specifically on prep. stranding but from a very different point of view (she was in touch with you the other day), could certainly give you exx of double preps from early and late Modern English.

[Nuria Yáñez-Bouza wrote:

Here are some examples from early Modern English.

With the same preposition:

Behinde the Lunges, towarde the Spondels, passeth Mire or Isofagus, of whom it is spoken of in the Anatomie of the necke (Helsinki Corpus, science, Thomas Vicary 1548, s2,p62,chVIII)

With different preposition:

Furthermore, Sir, if it please you to understand of the great unkindnes that my grandam hath showed unto me now latly, as the bringer herof can more planly shew you by muth, to whom I besech you to take credence on. (Corpus of Early English Correspondence, Germayn Poley, 1503, Letter CXLIV, p.179)

And one which I find particularly peculiar (and interesting to me, I've done a bit of work on that) is the combination of a stranded preposition with the where+preposition compound (e.g. whereat, whereby, wherefrom), instead of the
pied-piped preposition:

But if ye will sel it, send word to your son what ye will doe, for I know nothing els wherwith to help you with. (Helsinki Corpus, private letters, Isabel Plumpton, 1500-70, s2,L.162,p.199)

Nuria notes that she found no examples in her late Modern English corpus, which is "a selection of genres from ARCHER (1700-1900)", but she supplies these late Modern English gems from Jespersen (1909-49:III.10.5.1-3):

a pamphlet of which he came into possession of in London (Sharp, Browning 120)

a young Irishman, with whom I was once intimate, and had spent long nights walking and talking with (Stev.A.141)

She flags as especially idiosyncratic an example from Defoe, "where he strands the same preposition (to) and a different one (upon)":

I had nobody to whom I could in confidence commit the secrecy of my circumstances to, and could depend upon for their secrecy (Defoe)

I've added boldface to her examples to highlight the double prepositions.

Meanwhile, Patrick McCormick found a couple of additional examples from contemporary English:

The reason why they should be used is that not all databases provide sufficient information in the JDBC result set object to determine the table to which a column belongs to.

The group to which arthropleura belongs to dates all the way back to 420 mya and could very well have been the first arthropods to live on land more than in ...

The trouble with examples like these is that they may come from editing errors, where the author (for example) writes "the table that a columns belongs to", and then on a subsequent editing pass decides not to strand the preposition, and so changes "that" to "to which", but forgets to remove the other copy of "to".

But then again, maybe not.

Andrew Fader notes:

A big example of the double preposition is the phrase, "world in which we live in," which has a ton of Google hits, probably because of the Paul McCartney song "Live And Let Die" which features it. Some people actually think it's "world in which we're livin'," but it doesn't much sound like that.

Philip Spaelti sent in the same observation. I naively always thought it was "world in which we're livin'", myself, and never realized until now that the other construal was out there. But I should have, because (as Ben Zimmer just pointed out to me), Tim Warner took an extended look at this one on Mother Tongue Annoyances recently ("Worst Lyrics Ever? In Defense of Sir Paul", 3/31/2007), and Languagehat offered an earlier opinion ("Koaga and Wordtheque", 9/30/2005) that is less charitable to the "world in which we're livin'" interpretation.

I haven't listened to him performing the song recently. But in an r-less dialect, it's going to be just about impossible to distinguish unstressed "we" from "we're", especially before [l], where the inevitable formant transitions from [i] to [l] will pass through the same region of the vowel space as the schwa that would be the expected reflex of /r/ in that context; and "live in" is guaranteed to be homophonous with "livin'" in pretty all versions of English that have "g-dropping".]

Posted by Mark Liberman at May 4, 2007 07:09 AM