LINGUIST List 4.445

Wed 09 Jun 1993

Sum: Instrumental Subjects

Editor for this issue: <>

Directory

1. , Summary: Instrument Subjects

I recently posted a query as to why certain verbs in English and other
languages permitted agentive verbs to have instruments as subjects, while other
agentive verbs did not, and moreover, why languages differed as to which verbs
permitted this substitution. As examples I cited:
 He opened the door with the passkey.
 The passkey opened the door.
 German: Ich oeffnete die Tuer mit dem Schluessel.
 *Der Schluessel oeffnete die Tuer.
I received several replies, most of which raised problems of one kind or
another. Some questioned my acceptability ratings for the German sentences.
Perhaps native intuition plays a role here (I learned my German as an adult),
but the larger problem was one that others mentioned for English: that context
plays a very large role. Given the proper context, some otherwise implausible
sentences don't sound half bad.
Though context is important, it can't explain everything. In particular, it
can't explain why it's normal to say
 _This key can open the lock_.
but not
 *This sledgehammer will demolish the doghouse for you.
Even if we can find a context in which this last might be uttered, or when we
might say _Der Schluessel oeffnete die Tuer_, the we still need to explain why
a special context is necessary.
Some other respondents provided helpful data on Dutch (which behaves
approximately like German), Japanese, Hindi, and Kannada. I'm particularly
grateful for the last three and would be delighted to receive data on more
languages.
Once again, thanks to all who replied.
--Leo Connolly
*---------------------Messages Here-------------------------------------
Burkhard Leuschner <burkhard.leuschnerextern.uni-ulm.de> wrote:
While I might actually use this sentence to say that the key fitted
the lock
 *Der Schluessel oeffnete die Tuer.
 'The key opened the door.'
I would never use the second one
 Der Hammer zerschlug den Vogelkaefig.
 'The hammer smashed the birdcage.'
except in a very special situation, say in a story about Thor's
hammer, which sort of has a will of its own and might be dancing
around the cage and all of a sudden decide to dive down and smash it.
Bart Mathias <mathiasuhunix.uhcc.Hawaii.Edu> wrote:
I don't know anything about German cases (I'll hazard a remark on Japanese, my
area of "expertise," below), but it doesn't take natural forces (as I
understand the term) to demolish doghouses. E.g. "bomb," runaway truck," and
the like work for me. Further, I don't think it is a matter of nou class at
all. I can make your sledgehammer work in "He didn't get anywhere banging on
it with a screwdriver, but a/the sledgehammer demolished the doghouse."
 The difference between keys and sledgehammers may be that keys are closely
enough associated with opening locks that establishemnt of the real power in
context is unncessary, where the association of sledgehammers and doghouses is
a little too random.
Japanese allows natural forces such as winds, sunshine, to "do" somethings, but
I'm pretty sure the "key opens" case is out, along with the "bomb/runaway
truck/sledge hamme demolishes" cases; but I keep encountering things in
Japanese that look awfully English, so don't take that as gospel.
 I notice that I am comfortable with 'kaze/ookaze' "wind/gale" "knocked the
doghouse over," but if I (non-native speaker that I am) were going to translate
"the tornado demolished the doghouse," using a sino-Japanese word for
"demolish," I'd go with so-called "instrumental case" and passive, even though
Japanese doesn't like inanimate subjects with passives all that much either.
Sabine Bergler <berglercs.concordia.ca> wrote, in part:
Wouldn't "The sledgehammer demolished the doghouse faster than the crowbar
did." work and the starred sentence therefore be a context problem rather than
a grammatical one? This is just an aside.
 >
 >This example cannot be resolved by claiming that the subject must be
 >animate, since (at least) natural force subjects are possible.
 >
 > The tornado demolished the doghouse.
This is a real objection: "The tornado" here is not an instrument in the sense
of the sledgehammer. You can see this in the active:
 The doghouse was demolished with/*by a sledgehammer.
 The doghouse was demolished *with/by a tornado.
Thus tornado is always the subject, albeit inanimate and therefore called an
"instrument of distruction". Grammatically not the same.
 >Further, some languages permit fewer instrument-subjects than English
 >does. German _oeffnen_ 'open' does not permit them.
 >
 > Hans oeffnete die Tuer mit dem Schluessel.
 > 'John opened the door with the key.'
(aside: Notice that you'd say 'John opened the door by key.' in English...
Compare also "Death by sledgehammer" etc.)
 >
 > *Der Schluessel oeffnete die Tuer.
 > 'The key opened the door.'
Again, I see an out-of-context problem here. I am a native German speaker, and
I do not consider the first sentence natural but don't find the second one
much more objectionable. On the other hand, both the following are FINE for me:
 Hans oeffnete die linke Tuer mit dem kleinen Schluessel.
 Der kleine Schluessel oeffnete die linke Tuer.
The difference is that you would never mention the key unless you wanted to
contrast it or refer to some special property etc. "die linke Tuer" and "der
kleine Schluessel" set up enough of a possible context to license the explicit
mentioning of the key. Many "ungrammaticality" judgements are really due to
lacking context, in my opinion. Another reason might be that we do have a verb
that lexically entails a key, "aufschliessen", come to think of it, against my
pretheoretic expectations I can accept as a well-formed sentence: "Der
Schluessel schloss die Tuer auf."
 >But there is no general prohibition against them in German:
 >I believe _zerschlagen_ 'smash down' permits an instrument subject.
 >
 > Sie zerschlug den Vogelkaefig mit dem Hammer.
 > 'She smashed the birdcage with the hammer.'
 >
 > Der Hammer zerschlug den Vogelkaefig.
 > 'The hammer smashed the birdcage.'
This goes against my intuitions. The first sentence is fine, the second really
implies that the hammer was stored above the cage and fell and accidentally
smashed the cage (barring context, once again.) Of course, the following is
not as marked as the above:
 Der Hammer zerschlug den Kaefig in einem Schlag.
Jan-Wouter Zwart <zwartlet.rug.nl> wrote that the equivalent sentences in
Dutch are OK:
 (i) a. Jan opende de deur met de sleutel
 b. De sleutel opende de deur
He said that they get better when a context is added.
 (ii) a. Deze sleutel opent die deur/alle deuren
 b. Een van de sleutels zal de deur toch wel openen
Linda Shockey <LLSSHKEYcms.am.cc.reading.ac.uk> wrote:
... i can't see what's wrong with your starred sentence. if you add some extra
bells and whistles, it gets even better:
 the icepick didn't work, but a quick trip to the toolshed produced a
 remedy: the sledgehammer demolished the doghouse quite efficiently.
there's an understanding that someone is using it, but that's the same with the
key example, isn't it?
Richard Wojcik <rwojcikgrace.rt.cs.boeing.com> wrote:
You'll get people who claim that the "sledgehammer demolished" sentence is
perfectly ok. I would argue, in fact, that everyone who stars the sentence
probably would say it in an appropriate context. The problem here seems to be
a phenomenon that Jim McCawley pointed out many years ago--that linguistic
intuitions are bound to contexts. Whether or not you consider a string of
words to be ungrammatical depends on whether you can imagine an appropriate
situation to embed them in. That is why "the bomb demolished the house" is
fine, even though bombs are not forces of nature. Your difficulty in imagining
how a sledgehammer could demolish a house all by itself is the root of the
ill-formedness, not a restriction on the animacy of the subject. (In any case,
there is still the "courtroom witness" statement, where "this gun murdered the
victim" is a perfectly reasonable thing to say.) Agents and forces of nature
fit easily into subject slots because we have no difficulty in imagining their
ability to bring about results all by themselves.
Paul Black <Paul_Blackpost.ntu.edu.au> wrote:
 Sledgehammers and doghouses
Why can't you say 'The sledgehammer demolished the doghouse.'? Isn't it just a
matter of finding a meaningful context?
 Well, I was supposed to take the doghouse apart, but all I had was a
 claw hammer and a sledgehammer. I tried pulling out the nails with
 the claw hammer, but most of them wouldn't come, and I just got more
 and more impatient. The claw hammer just wasn't doing the job, so in
 the end the sledgehammer demolished the doghouse.
A little forced? Play with it:
 Can that key open this lock?
 Can that sledgehammer demolish that safe?
Shashi Nanjundaiah wrote:
Between Indian languages, there is a sort of dichotomy in the use of
instrument-as-subject. In Hindi, an Indo-Aryan language and a relatively recent
one at that, its use would be selective.
 usne kursi ko hathaude se tood daali
 "He/she broke the chair with a hammer."
While this is in the most common usage, it is unacceptable to say
 * hathaude ne kursi ko tood daali
 "The hammer broke the chair."
It is, however, not "weird" to say
 tuufaan ne peed ko ukhaad daala
 "The hurricane uprooted the tree."
In Kannada, the South Indian (Dravidian) language whose literature dates back
to 9th century, it is common to say
 avanu kattiyinda hannannu kattarisidanu
 "He cut the fruit with a knife."
but unacceptable to say
 * kattiyu hannannu kattarisitu
 "The knife cut the fruit."
Where Kannada differs from either Hindi or English (and predictably a host
of others) is in the next sentence:
 * candamaarutavu maravannu urulisitu
 "The hurricane uprooted the tree."
The unacceptability of this sentence gives rise to a change in voice:
 candamaarutadinda maravu urulitu
 "The tree was uprooted by/as a result of the hurricane."
As a student of sociolinguistics, I am tempted to think that there is a role
of the modernity of the language and degree of borrowed syntax at play here.
Although Kannada is a heavy-borrowing language, most of the borrowing is
lexical and not syntactic. Its syntax is quite conservative, really. This
is the case with other old Dravidian languages as well, so much so that
several times problems are encountered while code-switching between them
and, say, English (switching is very heavy in Kannada, Tamil and Telugu).
Hindi, on the other hand, has grown up borrowing heavily in lexicon and syntax
not only from other languages of its own group, but from other Indo-European
ones as well (mainly English).
I wonder whether the inanimate nature of the instrument has any part in this
mostly social--and less grammatical--taboo against instrument-as-subject.