About Rationally Speaking

Rationally Speaking is a blog maintained by Prof. Massimo Pigliucci, a philosopher at the City University of New York. The blog reflects the Enlightenment figure Marquis de Condorcet's idea of what a public intellectual (yes, we know, that's such a bad word) ought to be: someone who devotes himself to "the tracking down of prejudices in the hiding places where priests, the schools, the government, and all long-established institutions had gathered and protected them." You're welcome. Please notice that the contents of this blog can be reprinted under the standard Creative Commons license.

Saturday, June 03, 2006

Mona Lisa speaks?

I tried to stay away from any story related to the silliness surrounding The Da Vinci Code (ok, want my opinion? A mediocre novel, peppered with gross historical inaccuracies, certainly not a treat to the Christian faith -- unfortunately). But the news that a group of Japanese scientists has made the Mona Lisa (the subject of the famous Da Vinci painting by the same name housed in the Louvre museum) talk is really too much. It's bad science, and it's irrelevant science.

The idea is that one can go from an analysis of the two-dimensional image in the painting to a three-dimensional reconstruction of her skull, and from there to a computer-assisted simulation of her voice, since there is obviously a relationship between the exact positions of the mouth parts and the sound of the human voice. Indeed, the Japanese scientist who led the group claims that the process is “90% accurate,” though it would be nice to know how on earth he got that number, especially considering that this is certainly one hypothesis that cannot be tested (heck, we don't even know who Lisa was!).

It is bad science because although it is certainly true that the sound of our voice depends on the characteristics of our skulls (as well as, of course, the cultural influences we are subjected to throughout our lives, not to mention a variety of imponderables that are difficult to measure, for example the possibility of getting a bacterial infection that attacks our vocal cords), it simply doesn't follow that one can scan a painting and obtain a reliable estimate of what that particular human's voice sounded like. Going from a painting to a 3-D reconstruction introduces a variety of errors due to the approximation of the estimates, and these errors are compounded when one moves from the 3-D reconstruction to vocalization, because we don't know anything about the exact internal positioning and shape of crucial structures such as the voice box. I bet dollar to donut that the “accuracy” of such reconstructions is far less than the touted 90%.

This is also an example of entirely irrelevant science: who cares what the Lisa sounded like? What insight do we gain about the natural world, the human condition, or even the art of Da Vinci? The whole exercise becomes a farce when one realizes that the Japanese team's first attempt was to have Lisa speak, well, in Japanese! And even now, she speaks in modern Italian, not Renaissance Florentine (or was it Milanese?). This sort of crap certainly makes headlines on the tail of Brown's mediocre novel, but it doesn't do science itself (or the public) any favor.

Besides, what if Leonardo painted a mute?


  1. But how could it not be good science? I mean there are numbers and everything! :-)

  2. Now she can record an album. Maybe she can even compete on the next American Idol!

  3. Well did the Japanese study describe where the 90% figure came from? I would think that one easy way of coming up with such a number would be to go through the same process with 2-D paintings of currently living people and compare the results with recordings of their voices.
    Of course cultural and environmental factors will come into play, but they might be able to come up with a "ballpark" voice.

  4. Indeed, what a waste of time and money. But it "worked", made headlines, we're talking about it... Whatever they gain from it, if anything.


  5. I have always questioned the accuracy of 'reconstructing' flesh and muscle mass over a bone. I am a sculptor. I no longer work in a 'representaional' manner, but I do understand how layers form over an armature -- far from an 'exact' science, if science at all. Do dinasaurs and egyptians look anything like the 'modeled' examples? How about the one of JC? Give me a skull, I can slap some fat and hair on it.

  6. I can imagine the tech being used for law enforcment. Especialy for catching criminals and searching for missing people. Process a picture and get both the person's voice sample and profile in 3D.
    This could bring the Japanese bilions.


Note: Only a member of this blog may post a comment.