Chomsky vs Skinner Paper

What contributions was Skinner noted for?

Contributions to study of animal behavior

Main Issue with Skinner’s Book

Giving a “functional analysis” of verbal behavior

Functional Analysis

Skinner’s term; The examination of how ling elements serve communicative purposes + interact across dif layers of grammar + social contexts

Identification of variables that control this behav + specification of how they interact to determine a particular verbal response

Controlling of variables to be described completely in terms of: (which all have clear meanings in animal experiments)

1. Stimulus

2. Reinforcement

3. Deprivation

How does Skinner try to explain Human Behav?

1. Stimulus

2. Reinforcement

3. Deprivation

All of which have clear meanings in animal experimentation but not cog sci

Goal of Skinner’s Book

Provide way to predict + control verbal behavior by observing + manipulating physical environment of the speaker

Thinks advances in animal sci helps lang bc verbal behav is fairly well understood

What does Skinner Limit himself to?

Observables; Input output relation

Function

Skinner’s term; Describes the causation of behav

How does Skinner limit how observables should be studied?

Should be studied as 1 of the following:

1. Product of inborn structure

2. Genetically det maturation

3. Past Experience

Issue with Skinner + Behaviorism

Anyone who wants to analyze causation of behav like this can only use the data available; Will try to describe function by specificizing response with the history of inputs

Skinner’s Thesis

External factors consisting of present stimulus + history of reinforcement are of overwhelming importance; General principles revealed by lab studies provide basis for understanding verbal behav

Chomsky’s Issue with Skinner’s Thesis

Its far from justified; Lab studies of reinforcement can only be applied to complex human behav in the most complex + superficial way

How does Chomsky say only way to eval Skinner’s Argument is?

To review the functional analysis ex + determine the precise character of concepts

2 Categories of Animal Resp according to Skinner

1. Respondents (purely reflex resp)

2. Operants (emitted resp for which no obvi stim is discovered with)

Respondents

Purely reflex response elicited by particular stim

2 Categories of Animal Resp according to Skinner

Operants

Emitted resp for which no obvious stim can be discovered; What Skinner is primarily concerned with

2 Categories of Animal Resp according to Skinner

Stimulus Discrimination

The ability to respond differently to various similar cues, reacting only to a specific target stimulus that signals reinforcement or a specific outcome

Ex. only presses bar when light flashes

Discriminated Operant

A behavior that happens more often in the presence of a specific environmental cue, called a discriminative stimulus, because that behavior has been rewarded in that context before

Response Differentiation

Process that reinforces specific variations of a behavior while putting other variations on extinction

Secondary Reinforcer

A stimulus that acquires reinforcing properties through repeated association with an already reinforcing primary stimulus (such as food)

Skinner considers $ + approval to be secondary reinforcers bc association with food, etc

What does Skinner say Secondary Reinforcers are? (ex)

Skinner asserts that human social incentives (money, approval, prestige) operate as secondary reinforcers

Chomsky’s Refute to what Skinner says Secondary Reinforcements are

Chomsky notes that while token-reinforcement has been demonstrated in chimpanzee token-exchange experiments, extending this simple paradigm to explain human social motivation is completely unproved and implausible

Drive

Another variable that can affect the rate of bar pressing operant

Broad Defs

Characterizing any physical event impinging on organism as stim + any part of organisms as resp

What happens if Skinner accepts the broad defs (characterizing any physical event impinging on organism as stim + any part of organisms be)?

Must consclude that behav has NOT been demonstrated to be lawful; Must attrib big influence on actual behav to ill def factors of attention, etc

If a stimulus is any physical event impinging on an organism and a response is any behavior, then human behavior is demonstrably not lawful, because actual actions are governed by internal factors like attention, volition, and caprice

What happens if Skinner accepts the Narrow defs (Behav is lawful by definition but this is of limited significance, since most of what the animal does will not just be considered behav)?

Behav is lawful by def, but this is of limited significance since most of what the animal does will NOT be considered behav

If stimuli and responses are restricted only to events related by smooth, reproducible curves (like bar-pressing in a Skinner box), then behavior is lawful by definition, but the theory covers almost nothing of real-world activity

What def of behav does Skinner use?

Narrow

Issue with Controlling Stim

The word stimulus has lost all objectivity, bc no longer part of outside word + are driven back into organism; Just a retreat into mentalistic psych

We dont know what the current stim is till they respond

Verbal Operant

A class of responses of identifiable form functionally related to one or more controlling environmental variables

Issue with Verbal Operant

Doesnt answer basic questions, like how to det what are controlling variables, etc; Skinner provides no objective criteria to determine where a unit begins or ends in a stream of speech, how many units occurred, or what constitutes "similarity" between responses

Response Strength

Basic dependent variable in Skinner’s functional analysis; Det by rate of omission during extinction

Skinner argues that this is the only variable that varies significantly in the expected direction under conditions relevant to learning processes

Issue with Response Strength

Not clear how the freq of a resp can be attrib to anything BUT the freq of occurrence of its controlling behav; Contradictory + unclear

What does Chomsky say about Skinner’s choice of “stim”, “resp”, “control” and “strength”?

That theyre just paraphrases for words to describe behav

Operation of Reinforcement

Presentation of a certain kind of stim in a temporal relation with either a stim or resp

Reinforcing Stim

By its power to produce the resulting change in strength; Some stim produce the change, some dont → classified as reinforcing or nonreinforcing

Issue with Operation of Reinforcement + Reinforcing Stim

Useless in discussion of real life behav UNLESS we can characterize the stim which they are reinforcing + the situations + conditions in which they are reinforcing

Law of Conditioning/Effect

If the occurrences of an operant is followed by the presence of a reinforcing stim, the strength is increased

For skinner, learning is just change in resp strength

How does Skinner use the term ‘reinfrocement’?

Very loosely; So much so that assertion that reinforcement is necessary for learning + continued availability of behav is empty

Issue with the Skinner’s Claim that All Verbal Behav is Acquired + Maintained in Strength through Reinforcement? (+ also conditioning)

The term “reinforce” has no clear context; Only functions as a cover term for any factor related to acquisition/maintenance of verbal behav

Most Common Characterization of Reinforcement that Skinner Rej + Issue

Drive reduction; If a drive is postulated on the basis of the fact that learning takes place, the claim that reinforcement is req for learning is empty

Many evidence challenges view that drive reduction is req for learning

Imprinting

Most striking evidence for innate predisposition of animal to learn in certain direction + react approp; Unrewarded learning

Resulting behav can be refined with reinforcement

Ex. Young songbirds learn the complex aud characteristics of their species song in early youth before they can produce the song themselves + young birds rapidly learn to prefer + follow specific moving objs during critical windows without any observable food reward

Is it difficult to accept that reinforcement is req for learning?

Yes

Is it unlikely that insights from animal behav have broadened scopes to verbal behav that reinforcement theory has failed to demonstrate?

Yes

How do kids acquire a maj of their verbal + nonverbal behav?

By casual observation + imitation of adults + other kids; So not true that they only learn via differential reinforcement by adults

Issue with argument that experience (not innate capacity) must be overwhelming dominant factor in lang acquisition bc child speaks lang of groups?

This is superficial bc its speculative → more is involved in sentence structure than just insertion of lexical items, must be innate to an extent

What consists of actual data for studying lang?

• Behav of speaker

• Behav of listener

• Behav of learner

Automatic Self Reinforcement

Internal or solitary activities (talking to oneself, thinking, writing fantasy, composing music) are maintained because the behavior automatically affects the speaker as a listener

Chomsky’s Issue with Auto Self Reinforcement

Chomsky argues this deprives "reinforcement" of all objective content, noting that under this logic, accidentally cutting one's finger would count as a self-reinforcing thought → ‘reinforcement’ becomes a ritual paraphrase for ‘wants’, ‘likes’, or ‘wishes’ without explanatory power

Latent Learning (Maze Experiments) Def

Rats exploring a maze without any food reward exhibited a dramatic drop in errors once a food reward was subsequently introduced → This proved that animals construct cognitive maps of their environment through unrewarded exploration, directly disproving the behaviorist claim that reinforcement is necessary for learning

Latent Learning (Maze Experiments) Results

Learning that takes place without explicit reinforcement or drive-reduction, becoming evident only when an incentive is introduced later

Proved that animals construct cognitive maps of their environment through unrewarded exploration, directly disproving the behaviorist claim that reinforcement is necessary for learning

Direct Brain Stim Experiment Def

Demonstrated that rats experience reinforcement via direct intracranial electrical brain stimulation

Showed that reward does not depend on satisfying a biological physiological drive, undermining drive-reduction theories of reinforcement

Direct Brain Stim Experiment Results

Showed that reward does not depend on satisfying a biological physiological drive, undermining drive-reduction theories of reinforcement

Thrush Study Def

Gaping response of nestling thrushes matures from simple nest-jarring triggers to specific visual head-shape configurations

Thrush Study Results

These complex responses to specific visual shapes mature genetically through nervous system development rather than through differential reinforcement

Do humans have an unrewarded drive?

Yes

Ex. Scholars and scientists working intensely on papers that no one reads or cares about demonstrate complex human behavior driven by intrinsic curiosity rather than external feedback

Skinner’s Claim on Making Predictions/Imparting Info

Making a predication or imparting information (e.g., saying "The telephone is out of order" or "Whales are mammals") merely brings behavior under the control of a new stimulus via simple conditioning

Chomsky’s Resp on Skinner’s Claim on Making Predictions/Imparting Info

Hearing "The telephone is out of order" produces wildly diverse, context-dependent actions (fixing it, rage, throwing it out, or trying it anyway) depending on individual motivation → Calling this "conditioning" or "stimulus control" is a meaningless play-acting at science

Autoclitics

Verbal resp that expresses relations, arrangement, or qualifications b/w key lexical items (nouns, verbs, adj)

Skinner’s Idea of Sentence Production

Sentence production occurs by selecting key lexical items first and then applying autoclitics to fit them into a grammatical frame

Chomsky’s Refute with Autoclitics + Skinner’s Idea on Sentence Production

1. Lack of Physical Frame: Phrases like "Sheep provide wool" contain no physical grammatical frame at all

2. Grammaticality vs. Word Class Sequence: "Colorless green ideas sleep furiously" and "Furiously sleep ideas green colorless" share identical sequence frames of word classes, yet native speakers instantly recognize only the former as a valid sentence

3. Deep vs. Surface Syntax: "Struggling artists can be a nuisance" and "Marking papers can be a nuisance" share identical surface frames, but possess completely different underlying syntactic structures (proven when substituting "can be" with "is" vs. "are")

4. Psycholinguistic Timing Evidence (Goldman-Eisler): Studies on speech hesitation pauses reveal that pauses occur before major lexical items (nouns and verbs), contradicting Skinner's claim that key lexical items are chosen prior to the grammatical frame

Does Chomsky think Complex Sequential Motor actions (like speech) can be explained by S-R reps?

No bc syntax req generalized internal integ patterns imposed on behav by central brain mechanisms

Grammar

A grammar functions abstractly as a deductive theory that enumerates the valid sentences of a language in the same way a formal logical system enumerates theorems

Obligatory Rules

Grammatical rules that must be applied to produce any well-formed utterance in the language

Optional Rules

Internal selective mechanisms that represent the speaker's choices and intentions in generating a specific utterance

What does Lang Acquisition of a Child req?

Language acquisition requires the child to construct an abstract generative grammar from limited and noisy data → Hints at Universal Grammar

Universal Grammar

Because all normal children accomplish this remarkable feat of theory construction rapidly, uniformly, and independently of general intelligence, human beings must possess an innate hypothesis-formulating capacity