I just finished reading an all-time classic of experiment psychology “Principles of Behavior” by Clark Hull (1943). In the book, the author tries to delineate the first principles of a scientific theory of behaviour, based on a capillary survey of (mainly) animal experiments 1/n 
The first chapter presents his epistemological manifesto in favour of a mathematical (logico-deductive) formalization. It reminded me of the preface of “Theory of games" (1944). Back then use of formal approaches to social sciences was presented as a necessary paradigm shift
In chapter II, we dive into the specifics of what an objective theory of behaviour should be. Of prime importance here is the introduction of the notions of intervening variables and observables, which illustrate how the standard narrative about behaviourists is false
The intervening variable is explicitly described as an inobservable theoretical construct necessary to explain the relationship between stimuli and responses: a "mental" state. We are far away from the picture of behaviourism usually given by the cognitive "revolutionaries"
Chapter III and IV summarize the then current notions of perception and action. It is functional to a scientific theory of behaviour because these represent the input and the output of the system. What also clearly emerges from here and is confirmed in the remainder of the...
... the book is the author is clearly persuaded of the fact that behavioural experiments and analysis are a logical way to gain access to the neurophysiological mechanism governing the brain (a position that I share and is, if not dominant, still alive...)
After quickly addressing innate behaviours (Ch. V), Chapter VI finally introduces the acquisition of receptor - effect connections and primary reinforcement, conceived as drive (D) or need reduction.
The notion of Drive is key in Hull's theory. It plays a double role in defining, we now would say, the reward function (in a way similar to that recently proposed by @MehdiKeramati4 and @BorisGutkin) and of a discriminative, contextual, stimulus
Several Chapters are then dedicated to establishing how "habit strength" (sHr, the key theoretical construct of his theory) is affected by several factors, such as the amount and the delay of reinforcement and the temporal relationship with the stimulus and...
... including the relation between sHr and the level of intrinsic motivation (drive=food deprivation) which, en passant, is illustrated with an honestly pretty cool data visualization (especially given that we are in 1943)
Later chapters are dedicated to more complex scenarios, such as generalization and "patterning". It is sobering to realize that the complexity of the experimental designs used back then is much higher compared to the simplified "dog" and "cat" examples we often use in teaching
A specific treatment is given to the origin of noise (he called it "oscillation) that is hypothesised to derive from intrinsic, random variations of the underlying physiological potentials (@valentinwyart?).
Noise is also pointed out as one of the main causes why formal theories of social sciences have been hard to develop (in addition to anthropomorphism). He concludes the book how it started, with a very eloquent defence of the role of mathematics in social sciences
• • •
Missing some Tweet in this thread? You can try to
force a refresh
