Nugget #2 on Man-Computer Symbiosis

My initial reading of J.C.R. Licklider’s 1960 paper “Man-Computer Symbiosis” was somewhat disappointing, because it seemed to me that this visionary paper did not hold up quite as well as Vannevar Bush’s “As We May Think”. I’m trying to identify exactly what caused this effect in my mind, but I think in part it’s due to a contrast in how the two men presented their ideas. Dr Bush tried to describe various future concepts by painting a picture with intermediate steps, saying “it’s sort of like this, but done differently in a way we don’t yet know”, while Dr Licklider painted a picture but didn’t emphasize going beyond his description. In addition, Dr Bush’s descriptions are accessible to most readers, while Dr Licklider’s are a bit more obscure, relying more heavily on his readers’ background knowledge.

One particular example from Licklider’s paper is his description of the “trie” data structure as a memory mechanism, in section V, subsection C. I am familiar with that data structure from my “Data Structures and Algorithms” courses, so I could somewhat follow the verbal description he provides. But I would not expect “normal” people to grasp it easily, especially because no graphic was included. The Wikipedia article at http://en.wikipedia.org/wiki/Trie contains a useful graphic: Trie from Wikipedia
and a Google image search for trie provides many other examples. (As an aside, another problem I have with the trie is that it only provides rapid access if you know how to spell the index or key term. It does NOT of itself effectively implement associative memory, although it could be used as one component of such a scheme.)

I also noted several oblique references to several items of which I know, but I would expect most people don’t. These include BBN, a key Boston-area company in the development of networking and other defense-related technologies, the “SAGE System”, by which I think he meant the SAGE Semi-Automated Ground Environment air defense network, the IBM 704, the first mass-produced computer with floating-point capabilities on which FORTRAN and LISP were originally implemented (and which was small enough to be sometimes used as a ‘single user’ machine,) and various researchers and organizations such as Gelernter, Lincoln Labs, Bell Telephone Laboratories, and so on.

The nugget I chose to comment upon consists of the second paragraph of the paper:

“Man-computer symbiosis” is a subclass of man-machine systems. There are many man-machine systems. At present, however, there are no man-computer symbioses…. The hope is that, in not too many years, human brains and computing machines will be coupled together very tightly, and that the resulting partnership will think as no human brain has ever thought and process data in a way not approached by the information-handling machines we know today.

The first aspect I wish to discuss is that of a “symbiosis”. The second is the number of people involved.

The definition of symbiosis given by Licklider in the first paragraph is quite important. A key phrase within the definition is the “living together…of two dissimilar organisms.” Note the words “living” and “organisms”. A key issue I have with Licklider’s concept is the idea of the computer being a living organism. I do not deny that it COULD be, but the description of the computer’s possible functions later in the paper do not quite require that the computer be alive. The Wikipedia description of Organism contains two interesting sentences: “In biology, an organism is any contiguous living system…”, and “There is a long tradition of defining organisms as self-organizing beings.” (Which begs the question of ‘what is a being?’) Also, “many sources propose definitions that exclude viruses and theoretically possible man-made non-organic life forms.” So I’m not alone in questioning the idea of a computer being a living organism.

Dr Licklider began to address my issue in section I.B by introducing and describing the ideas of a “mechanically extended man” and a “humanly extended machine”. He states that these are NOT symbiotic systems, but rather semi-automatic systems: systems that started out to be fully automatic but fell short of the goal. He also indicates that man-computer symbiosis is probably not the final state: later on electronic or chemical “machines” will outdo the human brain. (I might call those “pure” Artificial Intelligences, especially if they are self-aware.)

My hangup with the use of the term “symbiosis” in this case, is that the term suggests that both parties to the relationship are separately alive. Clearly the human is, but what about the computer? In section III.B, his description of the strengths and capabilities of the computer gloss over that point, and he ascribes attributes of independence to the computer without explicitly supporting it. “Symbiotic cooperation” is only cooperation if either party can choose NOT to cooperate. And back to section I.B, he explicitly says a “mechanically extended man” (in which the computer is an extension of the man and does not have a choice) is not what he is discussing.

Now I am muddying the discussion by introducing “choice” as an aspect to help me evaluate whether the computer is alive, which is clearly not valid. The fig tree (from the initial paragraph of Licklider’s paper) doesn’t have a choice in whether the insect larva inhabits it, but we agree that both tree and insect are alive, and are organisms. The networked SAGE computer exhibited many aspects of a multi-cellular organism, but no one would assert it was alive, and most would assert it is not a real organism. Can we prove it is not an organism as a theorem, rather than assert it as a postulate? I’m not quite sure how, as it depends heavily on the definition of “organism” we choose…

There’s more to discuss about whether the computer is a living organism, and whether that is required for symbiosis, and whether the computer is an organism at all. But let’s move on to the next point: how many humans are involved?

Licklider’s initial terms suggest there is one man and one machine in the symbiotic relationship, and most of the paper can be read cleanly with that concept in mind. However, note that in section V.A, he says “Any present-day large-scale computer is too fast and too costly for real-time cooperative thinking with one man. Clearly, for the sake of efficiency and economy, the computer must divide its time among many users.” In the next paragraph he describes hypothetical “thinking centers”, perhaps even networked together. So included but not explicitly stated in his concept is the idea that through the (networked) computers, multiple MEN interact in a symbiotic relationship. If person A is in symbiosis with machine M, as is person B, then indirectly persons A and B are also linked. If machine M is linked to machine N, then all symbiotes of M and N are symbiotes of each other.

This brings to mind aspects of Joe Haldeman‘s Forever War series, and especially described within Forever Peace, in which people remotely-operating battle robots using electronic “jacks” implanted in their heads become inextriciably linked with and sympathetic (or empathetic?) to the other persons within the computer-assisted (I’ll call it) “mind meld”.

The development of semiconductor computers, and the rapid reduction in size, weight, power, and cost that Moore’s law has brought, means that Licklider’s assumption that computers are too big and expensive not to share is no longer valid. We now are blessed with the situation in which there are multiple computers per person, rather than multiple persons per computer. Therefore “symbiosis” between one man and one (or more) computers is more practical than that of multiple people to one computer. If the symbiote computers are then linked through networking, that provides additional capability and possibilities that Licklider did not extrapolate in detail.

Further, hark back to Dr Bush’s memex concept: the memex was a tool for an individual, so now consider a person/memex pair. Instead of the memex, substitute a computer symbiote. Instead of exchanging photographic trails between memexes, network the symbiotes. Now we have a worldwide network of augmented humans. Very exciting!! (And I suggest keeping that thought in mind as we read Engelbart’s paper next week. I don’t know how close he comes to that idea, but I do know he emphasized aspects of using computers to facilitate collaboration and teamwork…)

P.S. Most of my links here are to Wikipedia, because I’m mostly trying to point at definitions for people not already familiar with what I mention. Wikipedia then serves as a convienient jumping-off point for further exploration of those words and topics for anyone with interest in learning more…

Ideas for future followup:

  • Dig deeper into what it takes for a computer to be an organism.
  • Dig deeper into whether a computer can be a symbiote.
  • Watch for and discuss Engelbart’s ideas for networking multiple instantiations of NLS/Augment, and for transferring data among instances.
  • See if other readers felt the same way about Dr Licklider’s paper not being quite as “visionary” as Dr Bush’s, and why.

Leave a Reply

Your email address will not be published. Required fields are marked *