The Turing test is a proposal for a test of a machine's capability to demonstrate thought. Described by Professor Alan Turing in the 1950 paper "Computing machinery and intelligence," it proceeds as follows: a human judge engages in a natural language conversation with two other parties, one a human and the other a machine; if the judge cannot reliably tell which is which, then the machine is said to pass the test. It is assumed that both the human and the machine try to appear human. [from Wikipedia]
In order to keep the test setting simple and universal (to explicitly test the linguistic capability of the machine instead of its ability to render words into audio), the conversation is usually limited to a text-only channel such as a teletype machine as Turing suggested or, more recently, IRC or instant messaging.
Thus, the Turing test takes as its objective measure the best-informed subjective impression of the human judge.
I propose an analogous test, which I will call the Doofus-Turing Test Notwithstanding the tradition of naming such proposals after their proponents (anticipating the doofus retort). The Doofus-Turing Test involves communication between a human judge and a human subject, not between a human and a machine. Like the Turing test, however, it takes as its objective measure the best-informed subjective impression of the human judge.
In other words: I am borrowing the notion from the AI field that it is useful to employ such a measure of intellect. Turing himself suggested several objections which could be made to his test, and I am not here arguing whether the Turing test holds up as a useful test or not against such arguments. I have observed, however, that computer scientists and AI researchers won't shut up about the Turing test, and neither shall I.
If a human judge's subjective impression is a useful construct in measuring machine intelligence, then perhaps it can also be employed to measure aspects of human intelligence. That is the extent of the analogy I draw.
So, what aspects of human intelligence do I imagine might be measured by the Doofus-Turing Test? I imagine that the human judge will rank the subject along a spectum of doofusheadedness. I shall call this measure the Doofus Quotient. I do not here propose the method of assigning an objective rank, nor the mathematical form of the index. That is to be worked out later, by actual smart people (once this proposal passes their Doofus-Turing Test).
The Doofus Quotient, or DQ, scores the depth and consistency of the subject's understanding. A subject who manifests a shallow, superficial or rote understanding of a topic under discussion would score higher, and especially if the topic is one on which the subject holds him or herself out as having specialized knowledge. Conversely, subjects who evince deep knowledge and whose remarks hold together logically would score lower (they would be less doofusheaded). High DQ scores for adherents (or spouters) of idea-sets lacking strong empirical and theoretical foundation; low DQ scores for practitioners of rigorous and intellectually honest analysis.
My premise is that a human judge with sufficient acuity in a given area can easily recognize others with a similar or greater depth of understanding; and conversely, the same human judge can readily sniff out spurious “experts.” The more data points, the better (as usual); but it does not take very many data points, I propose, to distill out a useful DQ. If the subject produces accurate, consistent responses to, say, five penetrating questions in a row, this tells the judge a lot. If the subject can only manage canned, superficial bullet-pointy soundbites in response to the same five questions, all the while maintaining the faรงade of specialized knowledge, the judge can also learn much about the subject. Even a subject who has a deep understanding in some areas may yet score high on the DQ scale, if the subject over-leverages this expertise by feigning expertise in other areas, for example. Indeed, such behavior may warrant a multiplier of some sort!
The Doofus-Turing Test looks for logical inconsistencies. If there are any pockets of irrationality in the subject's reasoning, it is a sign that there could be many similar pockets. The Doofus-Turing Test is attuned to the spouting of talking-point memes, divorced from factual or empirical basis. The Doofus-Turing Test exposes the deployment of rhetorical trickery tending to deflect attention from the merits of the subject at hand.
Someone who spends too much time writing a piece such as this one would score higher on the Doofus-Turing Test; someone who stops here and attends to other matters in life would tend to score lower.
___
[Update: someone who came back to edit this post likely just spiked his DQ.]
___
& test of machiavellianism salon.com 9.13.1999
On the subject of the DTT and the DQ, which results therefrom, I am puzzled by a population stratification bias. There are categories of people (let's select politicians as one subgroup) who clearly cluster around a high DQ. They are forced to speak to constituents, the debate floor, or their party often at times when they have nothing to say, or for that matter, they are rhetorically modifying a party line to overemphasize its strength. For instance, the republican party during the Lewinsky scandal continuously turned the debate on the floor, whatever the issue actually was at the time, to the notion of morality in this country. One might argue that the morality of an illicit affair and even the lying to cover said act up is dispicable. But, at what point does the act of a single man, even one so high in office as the President condemn the morality of an entire nation?
Herein lies the efficacy of artfully used dufusheadedness. By taking a single situation and extrapolating it to win political sway, we also have a form of high DQ. I view the DTT as having two sub categories. There is high DQ A, which is highly rhetorical debate with no substance or intellectual reference wherein the dufus is merely talking to pontificate. And, there is high DQ B where the dufus knows exactly how wrong and untruthful the use of rhetoric is but chooses to perpetuate it because rhetorical sound bites amplify his position. This is the most dangerous form of a dufus. The high DQ B dufus is the one who turns entire political platforms into a few sound bites (i.e. decline of family values, defeating terrorism, etc.). It is ironic that in many cases those dufuses have been caught violating their own "rhetorical moral code" when they are caught in a scandal (e.g. Mark Foley) or act to undermine their own rhetoric (Abu Graib, Walter Reed, and a lack of suitable armor on Hummers).
What is truly incredible is the ability of the high DQ B dufus to quickly spin a tale so as to marginalize the subject of a scandal and form a new rhetoric which draws attention away from it (i.e. new sound bites such as "the Democrats do not support our troops because they want to bring them home"). I submit that the high DQ A dufus is easy to find, and often easy to discredit, but the high DQ B dufus is much more careful about what, where, and who actually sends the message (case in point, who was hung out to dry in the Valerie Plame scandal, those who contemplated the leak or those who actually perpetrated it?) The high DQ A dufus is one who plagarizes a speech in his first run for the Presidency, manages to rebound to make another run more than a decade later, only to make another stupid comment about his competition (e.g. Joe Biden). This is someone who knows rhetoric works but does not know how to use it or even when he is using it.
I do not want to leave out low DQ A and B dufuses as well. You might suggest that if the DQ is low, then one is not a dufus. On the contrary, I would argue that a party who is up against another group of nearly pure high DQ dufuses and still cannot find a way to win sway must also be called to the carpet. This is the low DQ dufus A. That is someone who clearly references and understands the argument and can argue intelligently and with data against a high DQ dufus but often does not formulate a proper argument or is so lazy as to not realize they can powerfully use the tools of the high DQ dufus B against them. The only gem in this crowd is the low DQ dufus B, who are few in numbers. These are the class of people who gather data, cross reference it, form an opinion, and articulately display it in a form such as to totally discredit the opinion editorial nature of the high DQ dufus.
Alas, I search for this person in the political realm and find none. There are too many high DQ dufuses, particularly B's, who so quickly make the situation unenjoyable for the low DQ dufus B, so why would anyone worth the job actually ever choose to do it.
May 26, 2007 11:01 AM
Low-D QB wondertwin powers activate! Form of a network; shape of a phase transition.
June 12, 2007 11:11 AM