« Should We Move to Speed Dating at the Meat Market? | Main | First Twelve in the Box »
Monday, June 04, 2007
Do-It-Yourself Reverse Browsewrap
Inspired by Ian Ayres’s LiabiliT (back story here), I've added some text to my browser’s User-Agent string, so that every time I point my browser to a web site, my computer includes in its initial request this little statement:
By responding to this HTTP request, you accept legal responsibility for any resulting harm.
Like Ian with his T-shirt, I would be interested to hear what legal effect (if any) people think this statement has. Have I successfully bound the web site to a reverse-browsewrap contract?
Some theories of browsewrap would say that I have, which I take to mean that there is something deeply wrong with those theories. It is impossible for a web site operator to have its software parse and understand the natural-language meaning of arbitrary text sent to it by a browser. To say that the web site operator has “agreed” by sending a response would be like saying that I have “agreed” to a contract by nodding politely at someone talking to me in Esperanto. It’s simply not reasonable to infer an intent to accept from my—or the web site’s—actions.
Thoughts?
(Firefox users, if you’d like to do the same, point your browser to about:config, right-click, select New > String, enter “general.useragent.override” for the preference name, and type in your new text. Fuller explanation here. You can also install User Agent Switcher to automate the process, or RefControl if you’d rather override the referrer part of the HTTP request. Do note that messing about with the information your browser transmits has privacy implications.)
Posted by James Grimmelmann on June 4, 2007 at 11:33 AM in Information and Technology | Permalink
TrackBack
TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8341c6a7953ef00df351decee8833
Listed below are links to weblogs that reference Do-It-Yourself Reverse Browsewrap:
Comments
Interesting -- but I'm not sure how you'd be able to argue that the string you include in the request is somehow binding on the website operator, at least where there's no reason to suspect that the website is programmed to recognize it.
UCITA Section 112: "(b) An electronic agent manifests assent to a record or term if, after having an opportunity to review it, the electronic agent: (1) authenticates the record or term; or (2) engages in operations that in the circumstances indicate acceptance of the record or term."
"(e) . . . (2) An electronic agent has an opportunity to review a record or term only if it is made available in manner that would enable a reasonably configured electronic agent to react to the record or term."
Of course UCITA is only the law in Maryland and Virginia. But I don't think under the Restatement there's any manifestation of assent to your term, or any exchange of promises or performance (what is it you're agreeing to do?). Even if there's consideration, sending the web page is not in reference (probably) to your term, and therefore that term is not binding under Restatement 2d Contracts s 23: "It is essential to a bargain that each party manifest assent with reference to the manifestation of the other."
Posted by: Bruce Boyden | Jun 4, 2007 3:08:55 PM
Bruce, just to be clear on the implications, do you think there is any significant difference between my User-Agent string and the terms of service that most commercial websites employ? I would maintain that when I request web pages via my browser, I don't do so "with reference" to any of the website's terms of service. The lack of consideration cuts against enforcing either. And we'll just leave UCITA by the wayside, since the world pretty much has.
Posted by: James Grimmelmann | Jun 4, 2007 3:49:43 PM
I do think there's a difference -- many sites feature a "terms" link at the bottom of the page, and I suspect a lot of people (though probably not all) would at least know to look for them there, even though very few actually read them. That's different than your User-Agent string, which I doubt anyone setting up a website would be expecting in that field. It's the difference between terms for your new refrigerator that you don't read that are included in the refrigerator box, and terms for your refrigerator that you don't read that are inserted in your Yellow Pages (under "Pizza") in the middle of the night. The first may become part of the contract, but the second certainly doesn't.
I also think the consideration issue is different. The website owner agrees to provide you certain pages, but only if you agree to the terms. But in your request, what service are you performing or promising? To read the web page?
As for UCITA, I think it's useful in exactly these sorts of situations, even outside of Va. and Md., since it directly addresses the issue. Certainly 112(e)(2) seems pretty reasonable.
Anyway, I'm not saying browsewrap is necessarily enforceable. But the argument for browsewrap seems much stronger than the argument for your User-Agent string.
Posted by: Bruce Boyden | Jun 4, 2007 11:48:40 PM
I wonder if and to what extent appending messages as such is part of the HTTP standard. If they can be passed as part of it, then I think you reach the legal adhesion questions. How they might resolve I'm not sure, but if sending such messages is outside the standard then I think it makes it much more difficult to assume the website could possibly be bound by any terms you've snuck in. Then you're not talking about whether affirmative behavior by the site constitutes acceptance of your terms because essentially the site, by having received information beyond what the standard allowed, might not have even been "aware" of them.
Posted by: Cathy | Jun 4, 2007 11:51:29 PM
Cathy: idiosyncratic User-Agent occupy an interesting grey zone in the HTTP standard. They are allowed, so that a message containing one is not ill-formed. There are various statements about how the field should be formatted and what it should contain, but they don't use the magic protocol-spec words (SHOULD, MAY, in all-caps, etc.). I think you're right that this is, in effect, "not a place someone should be expected to look."
Posted by: James Grimmelmann | Jun 5, 2007 10:55:27 AM
Bruce: You have a point about people knowing that web sites have terms at the bottom of the page. A CLS critique could just go to town on the asymmetry here. Corporations putting legal terms on their web sites have constructed a body of legally-significant knowledge, just because they all keep doing it. Whereas, when I, a consumer and an individual, do something extremely similar, it's a nullity, because no one "knows" to look at my message. Talk about your exercises in validating existing power structures! If enough powerful institutions just go ahead and do something, the rest of us are now forced to along! (I think that this truth is lurking underneath a number of more supposedly formalist theories of browsewrap.)
I think the consideration issue is symmetrical, in the sense that the request and response are the two sides of the HTTP exchange. You need consideration on both sides to find a contract. If I provide no consideration for my request, that lack of consideration should equally doom any contract based on the response.
Posted by: James Grimmelmann | Jun 5, 2007 11:06:22 AM
"Whereas, when I, a consumer and an individual, do something extremely similar, it's a nullity, because no one 'knows' to look at my message." Well, it's a nullity until you can convince enough of your friends to start doing it, at which point it becomes reasonable to look there for terms. So what you need is to start a movement.
Posted by: Bruce Boyden | Jun 5, 2007 1:51:45 PM
"So what you need is to start a movement."
I'm trying, I'm trying.
Posted by: James Grimmelmann | Jun 5, 2007 2:04:16 PM
So did you basically just give a definition for a sort of common law law?
Posted by: Colin | Jun 5, 2007 4:09:59 PM
The comments to this entry are closed.
