An Assessment of the Evolution of Research and Systems - Man-Machine Dialogue: Design and Challenges - page 5

(which was assumed, as we will see in section 1.1.1). Various paths were then
taken in serious MMD system design: a path close to AI, with a focus on
interpretation and reasoning issues, and a path that consisted of enriching the
automatic speech recognition systems. Both paths with their two separate
communities [VIL 05, p. 47], have recently come together again and allowed
various consistent MMD systems to reach fruition. These are the systems we
will now present.
1.1.1. First motivations, first written systems
Can a machine think? In 1950, Alan Turing relaunched this question that
had recurred throughout technology's history: he substituted the question
“can a machine imitate a human?” and suggested a game, or test, based on
imitation, which became famous by the name of the Turing test. At first, the
imitation concerns a man and a woman: the test subject talks with a man and
a woman in turn, through machine-typed pieces of paper, without seeing or
knowing anything about his/her successive interlocutors. The man has to try
and pass for a woman, and the subject thus has to guess which one is the man
and which one is the woman. Then, without the subject's knowledge, the man
is replaced by a machine. If the subject cannot identify either of the
interlocutors, then the machine passed the Turing test. This game, created at a
time when it was impossible to program an MMD system, was the source of
innumerable discussions, of various and varied assertions on the nature of the
machine or of the human being. The interesting thing here is the challenge for
computer science: to program an MMD system that can be thought to be a
human being. Turing does not give us many hints as to how to achieve that
result. The description of the test is focused on experimental conditions and
does not address the importance of language and dialogue in this approach to
thought [TEL 09]. Nonetheless, there are competitions organized today (such
as the Loebner prize) inspired by the Turing test. The 1950s correspond to the
first research motivations for MMD, information seeking and NLP. We should
point out that the Atala association, created in 1959, was originally called
Association for the study and development of automatic translation and
applied linguistics (“Association pour l'étude et le développement de la
Traduction Automatique et de la Linguistique Appliquée” in French), and
then became the Association for Natural Language Processing (“Association
pour le Traitement Automatique des Langues”).