The entropy is a statistical parameter which measures, in a certain sense, how much information is produced on the average for each letter of a text in the language. If the language is translated into binary digits (0 or 1) in the most efficient way, the entropy // is the average number of binary digits required per letter of the original language. The redundancy, on the other hand, measures the amount of constraint imposed on a text in the language due to its statistical structure, e.g., in English the high frequency of the letter £, the strong tendency of H to follow T or of V to follow Q. It was estimated that when statistical effects extending over not more than eight letters are considered the entropy is roughly 2.3 bits per letter, the redundancy about 50 per cent. Since then a new method has been found for estimating these quantities, which is more sensitive and takes account of long range statistics, influences extending over phrases, sentences, etc.

It was named M achina speculatrix. "conditioned reflexes" brought the study The conditioned, 01' neutral, stimulus to which In M spheri-latrix we had:1 reflex. MACHINA SPECULATRIX, photographed by time cx-POSUI'E, is attracted by light in hutch at right. CONDITIONED REFLEX requires this arrangement of nerve cells. With this arrangement the model is reasonably docile.

Full text available for a fee. (The paper was first presented in March 1950 at the National Institute for Radio Engineers Convention in New York.)See also: summary slidesPhilosophical Magazine (Series 7) 41:256-275

An excellent place to start. In this article, Turing not only proposes the Imitation Game in its original form, but addresses nine different arguments against AI, including Goedel's theorem and consciousness. Several recent arguments against AI are variations on the ones Turing enumerates. 'I propose to consider the question, "Can machines think?" This should begin with definitions of the meaning of the terms "machine" and "think." The definitions might be framed so as to reflect so far as possible the normal use of the words, but this attitude is dangerous....The new form of the problem can be described in terms of a game which we call the "imitation game."' I.—COMPUTING MACHINERY AND INTELLIGENCE. Mind 59, p. 433-460 (PDF from Oxford University Press).

The possible ways in which machinery might be made to show intelligent behaviour are discussed. The analogy with the human brain is used as a guiding principle. It is pointed out that the potentialities of the human intelligence can only be realized if suitable education is provided. The investigation mainly centres round an analogous teaching process applied to machines. The idea of an unorganized machine is defined, and it is suggested that the infant human cortex is of this nature. Simple examples of such machines are given, and their education by means of rewards and punishments is discussed.

Oliver Selfridge in The Gardens of Learning wrote: "I have watched AI since its beginnings... In 1943, I was an undergraduate at the Massachusetts Institute of Technology (MIT) and met a man whom I was soon to be a roommate with. He was but three years older than I, and he was writing what I deem to be the first directed and solid piece of work in AI (McCulloch and Pitts 1943) His name was Walter Pitts, and he had teamed up with a neurophysiologist named Warren McCulloch, who was busy finding out how neurons worked (McCulloch and Pitts 1943).... Figure 1 shows a couple of examples of neural nets from this paper---the first AI paper ever." From the introduction to the Warren S. McCulloch Papers, American Philosophical Society.http://www.amphilsoc.org/mole/view?docId=ead/Mss.B.M139-ead.xml;query=;brand=defaultAlthough an important figure in the early development of computing, McCulloch's goal in research was as much to lay bare the foundations for how we think as it was to develop practical applications - or in other words, to develop an "experimental epistemology" with which to relate mind and brain. Perhaps the most significant work to emerge from this period of McCulloch's career was his landmark paper with Walter Pitts, "A Logical Calculus Immanent in Nervous Activity" ( Bulletin of Mathematical Biophysics 5 (1943): 115-133). The "Logical calculus" was an attempt to develop just that: a rigorous description of neural activity independent of resort to theories of a soul or mind. Together with McCulloch and Pitts' follow-up work, "How we know universals: The perception of auditory and visual forms" ( Bulletin of Mathematical Biophysics 9 (1947) 127-147), the "Logical calculus" provided a compact mathematical model for understanding neural relationships laying the groundwork for neural network theory and automata theory, and forming the ur-foundation of modern computation (through John Von Neumannn) and cybernetics. (See Marvin Minsky, Computation: Finite and Infinite Machines, Englewood Cliffs, NJ: Prentice-Hall, 1967, for a very readable treatment of the computational aspects of McCulloch/Pitts neurons.")Bulletin of Mathematical Biophysics, 5, 115–137

‘Tarski discovered interconnections between such diverse areas of mathematics as logic, algebra, set theory, and measure theory. He brought clarity and precision to the semantics of mathematical logic, and in so doing he legitimized semantic concepts, such as truth and definability, that had been stigmatized by the logical paradoxes … Tarski’s famous work on definitions of truth in formalized languages (1933-1935) gave the notion of satisfaction of a sentence in a structure for first-order logic, second-order logic, and so on. This work had a profound influence on philosophers concerned with mathematics, science, and linguistics’ (DSB). ‘Tarski’s main contribution of the decade was his definition of truth. He claimed to have found the essential components by 1929, and they were stated without proof in the short paper [Der Wahrheitsbegriff in den Sprachen der deduktiven Disziplinen, 1932] communicated to the Vienna Academy in January 1932 which Carnap had seen. The first long version appeared in Polish as a book in 1933 … Acknowledging the work of Leśniewski on semantic categories, Tarski began by pondering the definability of truth for natural languages, and decided against it, especially because of unavoidable paradoxes; he stated a version of the liar paradox due to Łukasiewicz based upon giving the sentence “c it not true” the name “c”. But he saw a chance for a definition in a formal language by distinguishing it from a “second language, called the metalanguage (which may contain the first as a part)” and belonging to a “second theory which we shall call the metatheory”. This is seemingly the origin of those names: Carnap, to whom “object language” is due, mistakenly credited himself with “metalanguage” much later. The distinction was essential to Tarski’s theory, since the truth was a property in the metalanguage of a sentence correctly expressing some state of affairs in the object language: “it is snowing” is a true sentence if and only if it is snowing”. ‘Making use of recursive definitions, Tarski constructed a predicate calculus for the metalanguage, imitating the structure of the one in the object language. In order to ease the use of recursion, he worked with sentential functions rather than sentences: “for all [objects] a, we have a satisfies the sentential function ‘x is white’ if and only if a is white”. The crucial property was “satisfaction of a sentential function by a sequence of objects” in some domain, for from it he defined truth for any formal language with a finite number of orders of semantic category in terms of satisfaction by any sub-sequence in that domain. The background influence of Principia mathematica was explicit in his analogy between categories and simple types, and maybe in his decision to work with sentential functions ... ‘Comparing Tarski with Gödel, some of his techniques, and the impossibility result, correlate with incompletability and numbering; hence he was anxious to emphasise the independence of his own work, pointedly so in his Vienna note. However, his proof allowed for denumerably infinite sequences, while Gödel’s was finitary. Another contrast lies in Russell’s understanding: Gödel’s theorem always escaped him, but Tarski’s definition was described in his Inquiry’ (Grattan-Guinness, The Search for Mathematical Roots pp. 551-553). Tarski’s work was translated into German in 1935, and into English in 1956. From Bernard Quaritch Ltd Catalogue 2012/5 (https://www.quaritch.com/books/tarski-alfred/poj%C4%99cie-prawdy-w-j%C4%99zykach-nauk-dedukcyjnych/S870/). See also: J. Łukasiewicz. The Concept of Truth in Formalized Languages (http://www.thatmarcusfamily.org/philosophy/Course_Websites/Readings/Tarski%20-%20The%20Concept%20of%20Truth%20in%20Formalized%20Languages.pdf).

All this might anyhow be called metaphysics; but it is regarded as logic when adduced as bearing on an unsolved problem, not simply as information interesting for its own sake. The only one of these which is a distinct science is evidently (2). THE UTILITY OF LOGIC That of (1) above and of (3) are evident: the interesting ones are (2) and (4).