There are two aspects of this article upon which I would like to comment. The first is the extraordinary presumption of consciousness that runs through the entire piece, and the second is the misconceived attack on dual-system theories that appears towards the end.
Newell & Shanks (N&S) hold a strong presumption of consciousness with which they view all evidence. We can understand this by analogy to the presumption of innocence enshrined in those criminal justice systems based upon English common law. A decision process is conscious until proven unconscious; the burden of proof lies with those wishing to argue for unconscious decision making, and they must prove their case beyond reasonable doubt. Take, for example, the work on multicue learning and judgment reviewed early in the article. As the authors acknowledge, there are a number of studies, including those conducted in my own laboratory, which show that people can learn to make such judgments without being able to describe the basis for their choices as explicit knowledge. But using their presumption of consciousness and a very strong set of criteria (N&S Table 1), they nevertheless argue that the case for any of the actual learning or decision making being unconscious is not proven.
I ask the reader for the moment to adopt the opposite presumption: that decision making is unconscious unless proven beyond reasonable doubt to be conscious. How much harder would it be for the advocates of conscious reasoning to prove their case on this evidence than it was for their opponents under the opposite presumption? Even taking the civil law criterion of “balance of probabilities,” would the evidence again not clearly favour the advocate of the unconscious? It would obviously be easier for all of us to place a much stronger burden of proof on our theoretical opponents than we apply to ourselves, but we would clearly need to have a very strong justification for so doing. None is provided by the authors. The only basis for their presumption that I can see is common sense or folk psychology, as the everyday belief in conscious decision making is indeed widespread. In support of this, they do in places state or imply a default “intuitive” position of conscious thought.
The definition of conscious decision making that appears implicit in the authors' writing is “that which is accessible to awareness as evidenced by verbal report.” On this basis, I would argue for the opposite presumption to theirs. First, it is clear that while the processing capacity of the brain is vast, the thought of which we appear to be aware or able to report is very limited. Second, and despite the authors' attempt to discredit it, there is much evidence that our self-knowledge is poor and our verbal reports of our mental processes most unreliable. Finally, it borders in the mystical (or at least Cartesian dualism) to think of consciousness as some kind of “mind stuff” that has powers of causation. All of our conscious experiences are a product of the brain, because there is nothing else they could be. A conscious decision is one of which we become (at some point) conscious, nothing more nor less (see Evans Reference Evans2010, Ch. 7).
Not only do the authors believe they have discredited the idea of unconscious thinking, but also that in the process that they have successfully attacked dual-system and dual-process theories of higher cognition. Such theories distinguish between two kinds of processing: Type 1, which is fast, is automatic, and has high processing capacity, and Type 2, which is slow, is deliberative, and has low processing capacity (also known as System 1 and 2; see, e.g., Evans Reference Evans2007; Reference Evans2008; Kahneman Reference Kahneman2011; Stanovich Reference Stanovich2011). It is true that the distinction between conscious and nonconscious processing has been emphasised by some social psychologists (e.g., Wilson Reference Wilson2002), but it is emphatically not the foundation for contemporary dual-process theories of reasoning and decision making. As a dual-process theorist, I have argued, in common with others, that the conscious/unconscious decision cannot be the basis for the dual-process distinction because it is too vague and fails to define the key central properties of dual processing (Evans & Stanovich Reference Evans and Stanovich2013). I should also point out that in spite of defending the validity of much of the research that N&S criticise here, I have in common with them critiqued unconscious thinking theory and other strong assertions of the powers of intuition (see Evans Reference Evans2010, Ch. 4). This is because dual-process theory confines powers of reflective reasoning – and with it the ability to deal with novel and difficult problems – to Type 2 processing.
The case for dual process is in fact based not on the conscious/unconscious distinction but on the claim that there are two forms of cognitive processing which have distinctive properties and which reflect the operation of distinct cognitive and neural systems. Most of these properties are merely typical correlates, and few are defining features (Evans & Stanovich Reference Evans and Stanovich2013). I agree with Stanovich that Type 2 processing is distinguished both by its cognitive resources (central working memory, correlation with measures of cognitive capacity) and by its ability to engage in cognitive decoupling and hypothetical thinking. The apparent link of dual-process theory with consciousness comes only from the fact that some of the items attended in working memory are available to verbal report. But using broader definitions of consciousness, I have argued in detail that both Type 1 and Type 2 thinking have aspects that are conscious as well as unconscious (Evans Reference Evans2010, Ch. 7).
In conclusion, not only do I reject the authors' presumption of conscious decision making, which I believe to be shakily founded on folk psychology, but I also contest their implication that the conscious/unconscious distinction is the basis for contemporary theories of dual processing in higher cognition.
There are two aspects of this article upon which I would like to comment. The first is the extraordinary presumption of consciousness that runs through the entire piece, and the second is the misconceived attack on dual-system theories that appears towards the end.
Newell & Shanks (N&S) hold a strong presumption of consciousness with which they view all evidence. We can understand this by analogy to the presumption of innocence enshrined in those criminal justice systems based upon English common law. A decision process is conscious until proven unconscious; the burden of proof lies with those wishing to argue for unconscious decision making, and they must prove their case beyond reasonable doubt. Take, for example, the work on multicue learning and judgment reviewed early in the article. As the authors acknowledge, there are a number of studies, including those conducted in my own laboratory, which show that people can learn to make such judgments without being able to describe the basis for their choices as explicit knowledge. But using their presumption of consciousness and a very strong set of criteria (N&S Table 1), they nevertheless argue that the case for any of the actual learning or decision making being unconscious is not proven.
I ask the reader for the moment to adopt the opposite presumption: that decision making is unconscious unless proven beyond reasonable doubt to be conscious. How much harder would it be for the advocates of conscious reasoning to prove their case on this evidence than it was for their opponents under the opposite presumption? Even taking the civil law criterion of “balance of probabilities,” would the evidence again not clearly favour the advocate of the unconscious? It would obviously be easier for all of us to place a much stronger burden of proof on our theoretical opponents than we apply to ourselves, but we would clearly need to have a very strong justification for so doing. None is provided by the authors. The only basis for their presumption that I can see is common sense or folk psychology, as the everyday belief in conscious decision making is indeed widespread. In support of this, they do in places state or imply a default “intuitive” position of conscious thought.
The definition of conscious decision making that appears implicit in the authors' writing is “that which is accessible to awareness as evidenced by verbal report.” On this basis, I would argue for the opposite presumption to theirs. First, it is clear that while the processing capacity of the brain is vast, the thought of which we appear to be aware or able to report is very limited. Second, and despite the authors' attempt to discredit it, there is much evidence that our self-knowledge is poor and our verbal reports of our mental processes most unreliable. Finally, it borders in the mystical (or at least Cartesian dualism) to think of consciousness as some kind of “mind stuff” that has powers of causation. All of our conscious experiences are a product of the brain, because there is nothing else they could be. A conscious decision is one of which we become (at some point) conscious, nothing more nor less (see Evans Reference Evans2010, Ch. 7).
Not only do the authors believe they have discredited the idea of unconscious thinking, but also that in the process that they have successfully attacked dual-system and dual-process theories of higher cognition. Such theories distinguish between two kinds of processing: Type 1, which is fast, is automatic, and has high processing capacity, and Type 2, which is slow, is deliberative, and has low processing capacity (also known as System 1 and 2; see, e.g., Evans Reference Evans2007; Reference Evans2008; Kahneman Reference Kahneman2011; Stanovich Reference Stanovich2011). It is true that the distinction between conscious and nonconscious processing has been emphasised by some social psychologists (e.g., Wilson Reference Wilson2002), but it is emphatically not the foundation for contemporary dual-process theories of reasoning and decision making. As a dual-process theorist, I have argued, in common with others, that the conscious/unconscious decision cannot be the basis for the dual-process distinction because it is too vague and fails to define the key central properties of dual processing (Evans & Stanovich Reference Evans and Stanovich2013). I should also point out that in spite of defending the validity of much of the research that N&S criticise here, I have in common with them critiqued unconscious thinking theory and other strong assertions of the powers of intuition (see Evans Reference Evans2010, Ch. 4). This is because dual-process theory confines powers of reflective reasoning – and with it the ability to deal with novel and difficult problems – to Type 2 processing.
The case for dual process is in fact based not on the conscious/unconscious distinction but on the claim that there are two forms of cognitive processing which have distinctive properties and which reflect the operation of distinct cognitive and neural systems. Most of these properties are merely typical correlates, and few are defining features (Evans & Stanovich Reference Evans and Stanovich2013). I agree with Stanovich that Type 2 processing is distinguished both by its cognitive resources (central working memory, correlation with measures of cognitive capacity) and by its ability to engage in cognitive decoupling and hypothetical thinking. The apparent link of dual-process theory with consciousness comes only from the fact that some of the items attended in working memory are available to verbal report. But using broader definitions of consciousness, I have argued in detail that both Type 1 and Type 2 thinking have aspects that are conscious as well as unconscious (Evans Reference Evans2010, Ch. 7).
In conclusion, not only do I reject the authors' presumption of conscious decision making, which I believe to be shakily founded on folk psychology, but I also contest their implication that the conscious/unconscious distinction is the basis for contemporary theories of dual processing in higher cognition.