« previous post | next post »

Margalit Fox, "John E. Karlin, Who Led the Way to All-Digit Dialing, Dies at 94", NYT 2/8/2013:

A generation ago, when the poetry of PEnnsylvania and BUtterfield was about to give way to telephone numbers in unpoetic strings, a critical question arose: Would people be able to remember all seven digits long enough to dial them?

And when, not long afterward, the dial gave way to push buttons, new questions arose: round buttons, or square? How big should they be? Most crucially, how should they be arrayed? In a circle? A rectangle? An arc?

For decades after World War II, these questions were studied by a group of social scientists and engineers in New Jersey led by one man, a Bell Labs industrial psychologist named John E. Karlin. [...]

It is not so much that Mr. Karlin trained midcentury Americans how to use the telephone. It is, rather, that by studying the psychological capabilities and limitations of ordinary people, he trained the telephone, then a rapidly proliferating but still fairly novel technology, to assume optimal form for use by midcentury Americans. [...]

John Elias Karlin was born in Johannesburg on Feb. 28, 1918, and reared nearby in Germiston, where his parents owned a grocery store and tearoom.

He earned a bachelor’s degree in philosophy, psychology and music, and a master’s degree in psychology, both from the University of Cape Town. Throughout his studies he was a violinist in the Cape Town Symphony Orchestra and the Cape Town String Quartet.

Moving to the United States, Mr. Karlin earned a Ph.D. from the University of Chicago in 1942. Afterward, he became a research associate at Harvard; he also studied electrical engineering there and at the Massachusetts Institute of Technology.

At Harvard, Mr. Karlin did research for the United States military on problems in psychoacoustics that were vital to the war effort — studying the ways, for instance, in which a bomber’s engine noise might distract its crew from their duties.

I didn't know that Karlin was an alumnus of the Harvard Psycho-Acoustics Laboratory. As James H. Capshew explains (Psychologists on the March: Science, Practice, and Professional Identity in America, 1999):

The largest university-based program of wartime psychological research took place at the Harvard Psycho-Acoustic Laboratory (PAL). Employing nearly fifty people, including some twenty Ph.D. psychologists, the laboratory dealt with the problems of voice communication in the mechanized cacophony of modern warfare. It was headed by S.S. "Smitty" Stevens (1906-1973), a specialist in auditory psychophysics trained at Harvard under Edwin Boring in the 1930s and retained as a member of the faculty. He proved to be a gifted research administrator as  PAL gained major funding from the physics unit (Division 17) of the National Defense Research Committee.

The Psycho-Acoustic Laboratory was set up in tandem with the Harvard Electro-Acoustic Laboratory, headed by physicist Leo L. Beranek. The laboratories made a coordinated attack on the psychological and physical problems of sound control. Both were begun in December 1940 under the auspices of the National Research Council's Sound Control Committee, chaired by MIT physicist P.M. Morse.

[Stevens] was granted tenure and promoted in 1944, after being told a few years earlier by President Conant that he could expect to remain an assistant professor permanently.

S.S. Stevens was the originator of Stevens' Power Law, and also, apparently, the inventor of the doughnut-style headphone socket. He was 34 years old in 1940, and he was skillful, lucky or both in choosing other young people for his lab. Or maybe the lab's environment turned more ordinary people into extraordinary ones.

I'll mention just two of the many other PAL alumni who went on to distinguished careers: George Miller and J.C.R. Licklider. You can get some sense of their stature and contributions from the results of Google Scholar searches.

Miller's listing leads with "The Magical Number Seven" and "WordNet", and continues (in citation-count order) through Plans and the Structure of Behavior, "An analysis of perceptual confusions", Language and Communication, "Finitary models of language users", "Psychology as a means of promoting human welfare", "Some psychological studies of grammar", "Introduction to the Formal Analysis of Natural Languages", and so forth.

Licklider's intellectual journey took him even further beyond psychoacoustics, though he made significant contributions there, e.g. "Effects of Differentiation, Integration, and Infinite Peak Clipping upon the Intelligibility of Speech" (1948), "The Influence of Interaural Phase Relations upon the Masking of Speech by White Noise" (1948), "The intelligibility of interrupted speech" (1950, co-authored with Miller), "A duplex theory of pitch perception" (1951), "Basic correlates of the auditory stimulus" (1951), and so forth.

His most important ideas were in the realm of "Man-computer symbiosis" (1960) and "The computer as a communications device" (1968).

And at (D)ARPA from 1962 to 1968, he funded (and largely inspired) key developments in computer operating systems, in artificial intelligence, and computer networking — especially Arpanet, which metastasized beyond the military research establishment and turned into the Internet.

No blog post discussing the Harvard Psycho-Acoustic Laboratory would be complete without a reference to the Harvard Sentences, a set of 1360 phonetically-balanced sentences originally developed for "articulation testing" (J.P. Egan, "Articulation testing methods — II", OSRD Report No. 3802, 1944). Here "articulation" means "intelligibility", a usage established by Harvey Fletcher at Bell Labs in the development of his "articulation index" in the 1920s and 1930s (Harvey Fletcher and Rogers Galt, "The perception of speech and its relation to telephony", JASA 1950). There is also some relationship between the "Harvard Sentences" and a set of sentences used previously at Bell Labs; but I'm not sure whether it's analogy, inclusion, or equivalence.



  1. Steve Reilly said,

    February 10, 2013 @ 9:28 am

    "No blog post discussing the Harvard Psycho-Acoustic Laboratory would be complete with a reference to the Harvard Sentences. . ."


    [(myl) Lapsus digitorum. Fixed now.]

  2. Garrett Wollman said,

    February 10, 2013 @ 11:50 am

    J.C.R. Licklider also succeeded Bob Fano as director of MIT's Project MAC. It was during his tenure that Minsky's AI group broke away from Project MAC to form the AI Lab, a split that was not healed until 2003. While at DARPA, Licklider provided the seed funding for Project MAC (a three-year, $3 million grant), which began in 1963. and (as the Computer Science and Artificial Intelligence Laboratory) will be celebrating its 50th anniversary this year. That initial grant was used to buy an IBM 7094, specially modified with time-sharing hardware, on which the initial development of Multics was done before it was self-hosting. Among the notable developments in that time were electronic mail and Jerry Saltzer's TYPSET/RUNOFF document formatting system.

    Full disclosure: I work there.

  3. Mary Bull said,

    February 10, 2013 @ 12:02 pm

    This post and Garrett Wollman's additional info in his comment are fascinating to me. I recently watched the U.S. PBS program "Silicon Valley" and am amazed at how many brilliant minds were needed for all of us to have the conveniences in communication by both phone and Internet that we enjoy today.

  4. RobertL said,

    February 10, 2013 @ 11:53 pm

    All those intellects, and we still ended up with a telephone keypad that is arranged differently to the keypad on a computer keyboard – 1 at upper left vs 1 at lower left.

  5. Mary Bull said,

    February 11, 2013 @ 10:08 am

    @RobertL said,

    "February 10, 2013 @ 11:53 pm
    All those intellects, and we still ended up with a telephone keypad that is arranged differently to the keypad on a computer keyboard – 1 at upper left vs 1 at lower left."

    How true!

  6. spherical said,

    February 11, 2013 @ 10:26 am

    Blame the computer industry for sticking to the superannuated cash register format, including the large Enter key.

  7. LHC said,

    February 11, 2013 @ 4:47 pm

    @RobertL: you should read the NYTimes article. I'm not trying to be snarky: it explains that the telephone keypad design was based on research that found that users dialed faster and more accurately with that arrangement.
    What's truly annoying is that all those other devices didn't adopt the better, faster, more reliable ATT keypad layout. But note that many devices (ATM's for instance), DID adopt the more reliable ATT layout.

  8. Philip Lawton said,

    February 18, 2013 @ 7:40 pm

    Should 5.10 of the Harvard Sentences you linked to say "Mesh wire…" not "Mesh mire…"?

  9. snow day serendipity | oookblog2 said,

    May 26, 2013 @ 1:53 pm

    [...] posting by Mark Liberman at LanguageLog about the Psychoacoustics Lab at Harvard nudged the memory cells and set off a train of associations. The PAL was located in the basement of [...]

RSS feed for comments on this post · TrackBack URI

Leave a Comment