|THE HISTORY OF HANDWRITING
It's a lost art, another casualty lying on the shoulder of the information superhighway: handwriting.
It supposedly tells you a lot about a person. But now we "write" in faxes or E-mail; people court each other, fall in love even sight unseen over the Internet, using perfectly formed, precisely spaced characters on generic computer screens.
Tamara Plakins Thornton shares her (mostly) complementary perspectives on handwriting with colleague Sargur Srihari, professor of computer science and lead researcher in developing an automated system for the U.S. Postal Service that reads handwritten addresses. Happily, they agree, handwriting is anything but dead.
UB Today: How did each of you become interested in handwriting?
Thornton: I used to say I got interested in handwriting because my own is so bad, but I'm really interested in the concept of individualism and, at the same time, the idea of conformity. With handwriting, on the one hand you have people who are trying to get you to conform to an external model and to be obedient to rules. On the other hand you have people who resist that, who say "I'm unique, I'm not conforming to any model."
Srihari: At CEDAR [Center of Excellence for Document Analysis and Recognition], we are looking at what is perhaps the opposite of your interest, to see what is common about handwriting. What I try to do is find out, irrespective of who wrote it, can we read it? Can we make sense of what the writer intended? To some extent, that is possible, but our system is able to read addresses only 30 percent of the time. We can read nice, clean handwriting not necessarily the prettiest. The handwriting that has all the nice flourishes which is where the writers may be expressing their individuality those are harder to read.
UB Today: Is handwriting getting worse?
Thornton: There is a huge amount of literature saying that handwriting has gone downhill. I'm kind of skeptical because if you look at the 1950s, which we remember now as the golden age of penmanship, there was a huge amount of literature saying that penmanship was going downhill because of the typewriter and the Dictaphone and the telephone.
Srihari: I know that kids don't hand in handwritten assignments anymore in school. They spend less time writing, so they get less practice.
Thornton: What's interesting now is the development of these handwriting fonts for computers. You send in samples of your handwriting and they create a personalized font. One of the companies doing this says, "We don't just take your individual letters and string them together; we can contextualize them so we can make the t in the middle of the sentence look the way yours does."
Tamara Plakins Thornton
Srihari: It's your personal font. People get tired of things that are all printed with a word processor. We like to see something a little bit more stylized. Many of these handwriting fonts are quite popular.
Thornton: Are they legible by your software?
Srihari: The post office has asked us to look into it. They would like to recommend what handwriting fonts their business mailers should use. We are doing a small experiment at CEDAR to run through several of them to see which of these fonts are better than others so that they can say to their mailers, "Please use this one so we can sort your mail more efficiently."
Thornton: Just go to www.signaturesoftware.com. That company says they don't just take a sample of your alphabet so that they know what your b looks like; rather, they have special software that can contextualize each letter, so that a b preceded by an a will look somewhat different than a b preceded by an e. They say this is, in fact, what happens in real handwriting, so they are probably dealing with some of the same issues you are.
Srihari: We cheat a lot! We're more like engineers trying to solve a problem as opposed to trying to really understand what the basic underlying science is. We cheat in the sense that we try and finesse the problem and come at it from all kinds of angles. For example, if we know that in this ZIP code there are only three street names that are valid, why bother trying to read the whole thing? We try to finesse the problem without really trying to read it and we figure that's how people read, too. You bring in a lot of knowledge when you're reading someone's personal letter to you, even if the handwriting is illegible. The same goes for doctors' prescriptions, too. I think we'd all be really ill with all kinds of bad medicines...
Thornton: ...if pharmacists didn't know what the prescriptions probably said.
Srihari: The pharmacist obviously has a very good idea of what the doctor means. That's where we come from at CEDAR. We try and avoid the actual problem of recognizing all the differences in handwriting. We use two approaches: holistic, where you look at the global shape of the writing; and analytic, where you look closely at each letter.
Thornton: That's interesting, because when handwriting analysis first developed, as the science of graphology, there were the same two approaches: holistic and analytic. In the late 1800s, the German school of graphology was holistic; they looked at the gestalt. Whereas the French were analytical, and they developed a whole vocabulary of visual signs. They would say, for example, that if your t bar slopes up, it means you're ambitious, and if you have a t bar that slopes down it means you lack self-confidence.
Srihari: Actually, I had a student do a master's project with me to see if we could develop an expert system that would do handwriting analysis as an expert would. We wrote a program that had all those rules we got them from books on handwriting analysis built in. Then we analyzed the handwriting and it came out exactly how the books said it would.
UB Today: Well, I guess the gist of all this is that handwriting's not dead.
Srihari: Not at all. The most natural way to communicate among humans is speech. But after speech, the next most natural way to communicate is with handwriting because it requires very basic technology, just a stick in the sand, or pen on paper. It doesn't require any specialized technology or equipment.
Thornton: People are not going to carry laptops around to do their shopping lists. It's just not going to happen. And we'll always be able to write well enough. I just think we sort of go nuts about wanting everyone to write beautifully. We expect people to learn to swim well enough so they can swim to shore, but we don't expect them all to be Olympic butterfly stroke medalists! As long as people write legibly enough, I think that's fine.
Ellen Goldbaum is senior science editor for University News Services.