Kentucky Court of Appeals: domain names are not gambling devices

« previous post | next post »

We have previously discussed the claim of the Attorney General of Kentucky and a trial court's acceptance of that claim, that the domain names of internet gambling sites are "gambling devices" and therefore subject to seizure under Kentucky law. I am pleased to report that the Kentucky Court of Appeals has ruled that domain names are not gambling devices and on this basis has issued an order of prohibition barring the trial court from enforcing its seizure order. The majority held that:

…it stretches credulity to conclude that a series of numbers, or Internet address, can be said to constitute a "machine or any mechanical or other device…[sic] designed and manufactured primarily for use in connection with gambling." We are thus convinced that the trial court clearly erred in concluding that the domain names can be construed to be gambling devices subject to forfeiture under KRS 528.100.

In his dissent, Judge Caperton argues as follows. First, that the user's home computer together with the gambling site's computer form a compound device; Second, that the IP address of the gambling site is a necessary part of this compound device, and therefore that the domain name is itself a device.

I believe that this argument fails for three reasons. First, it conflates domain names with IP addresses. An IP address is indeed necessary to connect to a computer, but a domain name is not. A domain name is an alias for an IP address, one that is useful because it is likely to be easier for human beings to remember, but quite unnecessary. If you know the IP address of a computer, you can connect to it without knowing its domain name.

The second problem is that what the argument establishes, if anything, is that domain names are part of a device. Judge Caperton jumps from that proposition to the conclusion that domain names are themselves devices without argument. This is not a valid syllogism. It is possible for something to be part of a device without itself being a device, assuming that "device" has its usual meaning of "machine designed for a purpose". A piece of sheet metal, for example, is a component of a computer case, but it is not itself a device as it is not a machine and is not designed for any particular purpose. Here, however, we encounter a potential problem, namely that Black's Law Dictionary, which Judge Caperton cites, defines "device" as "an apparatus or article of manufacture". This definition is much broader than the usual definition: a piece of cloth and a bar of soap are both articles of manufacture but we do not consider them to be "devices". If one accepts Black's definition, our piece of sheet metal is indeed a device since it is an article of manufacture. Even by Black's definition, however, I submit that a domain name is not a device because, though it may be manufactured (that is, created by human beings), it is not an article of manufacture since, by virtue of being immaterial, it is not an "article".

Moreover, it is necessary for the State of Kentucky to establish not merely that domain names are devices but that they are "gambling devices", that is, devices whose primary intended function is gambling. Here the argument from whole to part founders, since a component of a device designed for a particular purpose is not, in general, itself a device designed for that same purpose. A computer is a device designed for computing, but the capacitors in its power supply are not therefore computing devices.

I therefore find Judge Caperton's defense of the State of Kentucky's claim flawed for three reasons:

  1. it improperly conflates IP addresses, which are necessary, with domain names, which are not;
  2. it relies upon the inference from "X is a component of a device" to "X is a device", which is false under the usual definition of "device";
  3. it relies upon a false inference from "X is a component of a gambling device" to "X is a gambling device";


  1. Ryan Denzer-King said,

    January 24, 2009 @ 8:13 pm

    A victory for logic! How rare and unexpected.

  2. Nick Lamb said,

    January 24, 2009 @ 9:04 pm

    Hmm, "trail court" is presumably a typo, and not an innovative coining of some kind.

    I'm glad to see that this went as I'd hoped, but it's disappointing that Caperton's grasp of the subject he was ruling on is so poor. One of the nice things to see in some critical High Court and Law Lords decisions in the UK were that judges ensured they understood what was in principle actually happening, rather than relying on a series of analogies. Analogies are treacherous, and I think in this case analogy has betrayed at least one of the judges.

  3. Dan T. said,

    January 24, 2009 @ 9:29 pm

    If there are Circuit Courts, why not Trail Courts, where the judge hits the trail to get from one venue to another?

    Some (perhaps most) Web sites are set up as virtual domains, where many different sites at different domain addresses are on the same physical machine, often under the same IP address, so that there is no way to access a specific site other than by domain; if you just type in the IP, you get some other site on the same physical machine.

  4. Jeff Epler said,

    January 24, 2009 @ 10:04 pm

    In fact, the domain name of the website being accessed can be as important as the IP address in determining what content will be served by the user. Through the magic of "virtual hosting" defined in RFC2616, the client transmits to the server the name by which it knows the server, and the server may respond with different content depending on that name.

    My own webserver responds to no fewer than a half-dozen different names with differing content, and systems operated by big web-hosting providers may respond to hundreds or thousands of different names with different content.

    While I agree with the rest of your argument, I think you should consider skipping the DNS name / IP address distinction.

  5. Nick Lamb said,

    January 24, 2009 @ 10:49 pm

    Dan T. I think I mentioned that in one of the earlier comment threads for this series of blog entries. However, "seizing" the domain name still doesn't entirely help you, because in HTTP/1.1 the name is presented by the client. Taking the "Joe's Casino" sign down from outside doesn't stop people who know where to go from ringing the bell and asking for Joe.

    (Oh the irony of resorting to this analogy after my earlier comment)

  6. Wells Hansen said,

    January 25, 2009 @ 9:37 am

    As a classicist, not a lawyer, it appears to me reasonable to argue that if a component is part of a device, then the part is subject to seizure if the device as a whole is subject to seizure, provided that the part can be shown to be an integral part of the device. Is this not the reasoning behind the “compound device” argument? A sheet of metal in a computer case and a capacitor in a computer would reasonably be subject to seizure in a circumstance in which the computer is subject to seizure. This, I believe, is the logic behind the argument that a domain name is part of a “compound device”. If the DEA can seize a vehicle that has been used to transport illegal narcotics, it does not seem reasonable to argue that the DEA cannot seize the wheels, or any other part of the vehicle, for none of the parts is, in itself, a vehicle.

  7. Dan T. said,

    January 25, 2009 @ 3:41 pm

    The "machine" in question seems to be claimed to consist of the entire Internet, and all computers, connections, namespaces, and other things that make it up. It would be something of a stretch for the state to demand forefeiture of all of those things, including every computer in the world that is or has ever been connected to the Internet, even if it has never been used for gambling, but that is what would be encompassed if that concept were to be taken literally.

  8. Mr Punch said,

    January 25, 2009 @ 4:17 pm

    With regard to your second point: Courts in various states have ruled that book publishers qualify legally as "manufacturers" (e.g., for the purposes of certain tax credits) even if they do not actually print the books themselves; so presumably an intangible (a text, perhaps a domain name) can be an "article of manufacture."

  9. Jens Fiederer said,

    January 28, 2009 @ 1:03 pm

    Really, although your reasoning is faulty, it should be blatantly clear from the definition that the domain name is not a device.

    a domain name is not a device because, though it may be manufactured (that is, created by human beings), it is not an article of manufacture since, by virtue of being immaterial, it is not an "article".
    is absurd.

    Clearly, "the" is immaterial, and just as clearly, "the" is an "article."
    So much for your argument.

    However, unless the domain name was something in the nature of "the", "this", etc., your conclusion is nevertheless unassailable.

  10. No, they’re not “gambling devices” said,

    January 28, 2009 @ 4:12 pm

    […] An appeals court rules against Kentucky's seizure of 141 offshore-casino domain names [Randazza, Citizen Media Law; earlier here, here, and here] More (via comments): Bill Poser, Language Log. […]

  11. Rose said,

    January 29, 2009 @ 3:31 am

    Not to confuse the issue, but:

    As this is Language Log, I ought to remind you that Kentucky is not a state, but a commonwealth.

  12. Arnold Zwicky said,

    January 29, 2009 @ 8:14 am

    To Rose: Kentucky is a state, but it calls itself a "commonwealth" (as do Massachusetts, Pennsylvania, and Virginia). The U.S. has 50 states (not 46).

    There are two senses of state here, and it's not contradictory to say that four of the 50 states of the U.S. are commonwealths. Insisting that state is acceptable only in its narrower sense is just annoying.

RSS feed for comments on this post