November 06, 2006

From one non-proscriptivist to another

Bill Weinberg recently offered a linguistic argument against the use of "open source" as a verb ("Commentary: Open Source is not a verb", NewsForge, 11/4/2006):

I am a linguist by training. Long before I delved into free software and was snagged by the quagmire of marketing, I pondered the marvels of morphology, the grimness of grammar and the splendor of semantics. It is only natural then that my wrangling criticism of industry-speak, in both technical and literary modes, is informed by ingrained linguistic sensibilities, descriptive and proscriptive. Given my background, I find it vexing when open source is used as a verb.

In my travels with OSDL, I frequently hear our eponymous Open Source employed as a transitive verb. As in "My company open sourced our product." Now, I am no petty proscriptivist.

Me neither. But I'm happy to offer some free editorial advice to a fellow linguist.

First, I think you mean "prescriptivist", not "proscriptivist". The OED tells us that a prescriptivist is "An adherent or advocate of prescriptivism", and that prescriptivism is "The practice or advocacy of prescriptive grammar; the belief that the grammar of a language should lay down rules to which usage must conform". There's no OED entry for "proscriptivist", but we could regard is as a regular derivation from proscriptive, which is glossed as "Characterized by proscribing; tending to proscribe; of the nature or character of proscription". The verb proscribe has the glosses

I. 1. trans. To write in front; to prefix in writing. Obs. rare. Perhaps a scribal error for prescribe.
II. 2. To write up or publish the name of (a person) as condemned to death and confiscation of property; to put out of the protection of the law, to outlaw; to banish, exile. Also fig.
 b. To ostracize, to ‘send to Coventry’.
3. To reject, condemn, denounce (a thing) as useless or dangerous; to prohibit, interdict; to proclaim (a district or practice).

And the noun proscription is glossed as

1. The action of proscribing; the condition or fact of being proscribed; decree of condemnation to death or banishment; outlawry. Also fig.
2. Denunciation, interdiction, prohibition by authority; exclusion or rejection by public order.

"Word rage" may be common among English-language prescriptivists, but ostracizing, death, banishment and confiscation of property are not really on the agenda here. Rejecting, condemning, denouncing, prohibiting and interdicting might be, so it's not nonsensical to use "proscriptivist" to mean "someone who condemns or denounces others for misuse of words". But it's confusing to coin a new word, when there's an old one that's almost identical in sound and essentially equivalent in meaning. And you're likely to make a bad impression on your readers, who may suspect you having committed a malapropism or created an eggcorn. So my advice is to proscribe the use of "proscriptivist", and to prescribe "prescriptivist" instead.

Weinberg's commentary continues:

English is a dynamic, productive language in which nouns can become verbs, and verbs can return the favor. Consider the word source (n. from Middle English sours, from Anglo-French surse spring, source, from past participle of surdre to rise, spring forth, from Latin surgere). Today, source is as often uttered as a verb as it is a noun, as in the dreaded labor term, outsource.

Uh oh, the frequency illusion strikes again! As Arnold Zwicky put it, "once you've noticed a phenomenon, you think it happens a whole lot".  And as Arnold asked,

Why do people ... who propose to offer authoritative advice to educated people not use standard sources of information? ("You could look it up", as Casey Stengel is reported to have said, with reference to his claim that most people his age were dead.)

One way to "look it up" in this case would be check examples of the word source on the web. So I read through the first ten pages of Google's returns for a search on source, without finding any examples of source as a verb. This suggests to me that it's unlikely to be true that source the noun and source the verb are equally common these days.

But Weinberg wrote "uttered", so maybe we need to check conversational use. Well, in the 26 million words of English-language conversations indexed at LDC Online, there are 392 instances of the word source, of which 391 are nouns, and one is a verb:

now everybody that can out source to a cheaper you know find a cheaper worker somewhere

(By the way, Weinberg's sentence offers another classic example of the McKean/Skitt/Hartman Law of Prescriptive Retaliation. When he wrote that "source is as often uttered as a verb as it is a noun", I think he left out an "as". I could be wrong about this, because both wordings give me a sort of unpleasant headachy feeling, but I believe that the clause ought to read "source is as often uttered as a verb as it is as a noun." In any case, it's probably not a good idea to create a 14-word clause involving either three or four copies of the word as. A better option might be something like "source the verb is now as common as source the noun". It's still false, but it reads better.)

We haven't quite gotten to Weinberg's real point yet, but we're getting closer. He continues:

What I find nettling is the presumption of what syntacticians call agency. In pragmatic grammar (as opposed to case grammar), the subject of a transitive verb is the agent that performs some act upon the patient or direct object of the verb. Dog [agent] bites [verb] man [patient]. The dog bites the man because it wants to, because it can. (Maybe a better example for software is "Cat throws up hairball"). But is it meaningful to say that the owner of a piece of code can open source that code, by fiat?

This is confusing. Is he saying that all transitive verbs have agents as subjects? In all flavors of grammar that I'm familiar with, some transitive verbs have agentive subjects and some don't. Here are a few examples where transitive verbs have subjects that are causes or themes or experiencers, not agents:

A fallen tree blocked the road.
The noise bothered her.
The bullet entered his chest and lodged near the spine.
Everyone in the room heard the explosion.

In those sentences, "blocked", "bothered", "entered" and "heard" are perfectly good transitive verbs, although none of their subjects are agents by the usual linguistic definition. Certainly none of them "performs some act upon the patient ... because it wants to". But in any case, someone who makes software available under an open-source license is both legally and linguistically a sentient agent who intends the result that is achieved.

Well, let's put all the grammar aside, because we're about to get to the real point:

There are actually four distinct stages for source code, only one of which I consider open source. ... The first is source code as documentation ... The second is source code as bait ... The third is source code under an OSI license ... The fourth and canonical scenario that embodies the true meaning of open source is a community of developers and users cooperatively building, deploying and maintaining project code.

OK, fair enough. Weinberg's idea seems to be that we shouldn't say "(person or company) A open-sourced (software system) X" just to mean "A made X available under an open source license", because X won't really be true "open source software" until and unless it comes to have an active community of developers and users.

I agree with Weinberg that "Without community, the source code behind open source is just a dusty tome, lifeless, static and unread". But lifeless, static and unread open source software is still open source software.

And therefore it's not only meaningful, but also true, to say that the owner who released some code under an OSI license "open sourced" it. You can object to this usage on aesthetic grounds, if you want to, but the business about subjects and agency is beside the point. English syntax and and semantics are neutral on this one.

[Hat tip to Tiego Tresoldi.]

[Update -- Bill Poser writes:

Bill Weinberg seems to be confusing a canonical association between transitive subjecthood and agency with a rigid implication. There are languages that have something stronger than what English does. In Japanese this association is sufficiently strong that, as Susumu Kuno pointed out years ago, it is generally not possible for a transitive verb to have an inanimate subject. To say "History repeats itself", for example, is bizarre in Japanese.

If he were writing about Japanese, I would just have pointed out that sentences like "IBM open-sourced UIMA" in fact have perfectly good agentive (and even quasi-animate) subjects.

What he's really saying is that the real agent of the open-sourcing process is the developer and user community, not the software owner. The trouble is, that's a moral or political judgment, not a linguistic one.

The usage fact is that phrases like the following are often found:

Yahoo! made this library available under an open source license.
The Redmond company made WiX available under an open-source license on
...we made it available under an Open Source license available at [10].
The CMU Sphinx project ... has made Sphinx2 available on SourceForge under an Open Source license.

There's nothing wrong with these sentences, in terms of syntax, semantics or word usage. And it's also sanctioned by the norms of English to refer to software that is available under an open-source license as "open source". As a result, it would be perfectly natural, from the point of view of English morphology and syntax, to rephrase each of those sentences using the causative neologism to open-source <something>, meaning to make <something> open source. Thus "The CMU Sphinx project has open-sourced Sphinx2 on SourceForge".

You might decide against this rephrasing because you don't like to make a new causative verb out of a complex nominal of the form adjective+noun -- that would be a reasonable stylistic preference. But to reject such usage on the grounds that "it takes a village to open-source a program" (as we might paraphrase Weinberg's argument) confuses morphosyntax with politics.]

Posted by Mark Liberman at November 6, 2006 12:04 AM