Subject: [IFETS-DISCUSSION:176] Re: IFETS-DISCUSSION digest 18
From: Andrea Downes (bryher34@hotmail.com)
Date: Wed 30 Aug 2000 - 00:53:33 MEST
From: "Andrea Downes" <bryher34@hotmail.com> Subject: [IFETS-DISCUSSION:176] Re: IFETS-DISCUSSION digest 18 Date: Wed, 30 Aug 2000 10:53:33 +1200
Due care, indeed, must be taken in the use of words...
Bob Leamnson wrote,
> If the discussion on "interaction" is not to become quite fuzzy and
> nebulous it's important that the word not be misused. "Interacting" means
> that each of two entities has an effect on, or changes in some way, the
> other.
No. Interaction should be defined as "an exchange of information between
two or more entities." Such an exchange may result in a behavioral change,
but there is no requirement that it do so.
In the definition above, I use the term "information" in a precise sense.
Information is not merely data flow. It is data flow which has semantic
content to the receiver. Specifically, from the point of view of the receiver,
the number of possible states of affairs in the world is reduced when
infromation is received. See Fred Dretske's Knowledge and the Flow
of Information for a detailed description.
> While "interacting" with a book might be an expression to suggest a
> strong "reaction" to the author's thoughts, it is in fact a misuse of the
> word since the reader's thoughts and ideas do not in any way have an effect
> on the book.
One's "thoughts and ideas" do not have an effect on anything; we are not
(so far as I know) telekinetic. Interaction always acts through a medium,
whether it be via voice, typing, push button, or some other physical
transmission device.
In the case of a book, the degree of interaction is limited because
most of the semantic content in a book is static. One cannot change
the words as they are printed on the page. But a book does have
a number of variable semantic properties: where it is located, which
page it is open to, for example.
A book, prior to its interaction has no information about what an
agent wants to see. We call this state "closed". When a person
picks up a book and opens it to a specific page, the book now has
that information. We call this state "open". Because a book is stupid,
it cannot act on this information; nonetheless, the information has
been transmitted from the agent into the book.
Granted, the amount of information transmitted to a book is very
small, and though the book may transmit a great deal of information
to the reader, the interaction per se is nonethess small. But the act
of leafing through a book, paging forward, glancing at the index or
table of contents - all these constitute interacting with the book.
> So it is that a student can "react" to a website but can "interact"
> only with another living person. I am assuming that Brent intends
> "interaction" in its dictionary meaning.
There is no reason to suppose that a person can interact only with
a living person. This supposition is based on the idea that only a
person can contain semantic content. But all manner of entities may
contain semantic content. True, they are not aware of this content
in any self-reflective way, but self-reflection is not an essential
component of interaction.
Brent writes,
> Muhammad Betz related that:
> "Another angle related to learning and interaction is the principle of the
> immediacy of reinforcement or feedback, which is a behaviorist tenet. In
> Brent's introductory paper, he explains of the frustration of on-line
> learners when they do not receive timely feedback and are left waiting for
> responses from either instructors or other students. Is "interaction" or
> the perception of interaction related to this principle i.e., the presence
> or absence of feedback to questions or responses?"
Interaction only occurs when the flow of information occurs in more
than one direction. Thus, an email message submitted and ignored
does not constitute interaction with the recipient, because the recipient
never replied. Insofar as interaction is desired by the student, this
lack of interaction is frustrating.
The term "feedback" denotes a specific semantic relevance to the
original transmission of information. It is not clear that feedback
per se would be required from the student (though one suspects
that information relevant to his previous submission would be
preferred to information not relevant to his previous submission,
though one would also suspect that some information, relevant or
not, would be preferable to none at all. This could be measured
empirically).
--Stephen Downes downes@ualberta.;ca
--
p.s.
In reading through the other posts I see an emphasis on human interaction. But because interaction is constituted of the exchange of information, I would think that the modality of interaction (specifically, human vs. non-human) is not a factor inherent in the quality of the interaction.
I have already alluded to one indicator of quality: semantic relevance to the content of the previous transmission. Other writers have discussed elements such as the level of academic discourse or the personal nature of communications as being other indicators.
As of this writing, no non-human is capable of quality in the terms defined above. Certainly, while a computer can react in some way to an individual message, a computer's grasp of content is often hit and miss; its responses are often inappropriate. And computers are for the most part not capable of creative thought or higher-level academic discourse.
But we can imagine a world in which computers do have these capacities. One wonders, then, whether in such a world a human component would be an essential element of high quality interaction. Could I obtain as much satisfaction interacting with a skilled, creative and sensitive computer as I would with a similar human. There seem to be no a priori reasons to think not.
______________________________________________________________ You are subscribed to the IFETS-DISCUSSION mailing list. For more information go to http://www.valdosta.edu/listproc/ ______________________________________________________________
This archive was generated by hypermail 2a24 : Wed 30 Aug 2000 - 00:58:46 MEST