Joshua Knobe has added considerably to our knowledge of the ways in which ordinary people attribute intentions and make judgments regarding causation. In this commentary, I do not want to criticize his claim that the competencies agents deploy in making these judgments are deeply suffused with normative influences. However, I will suggest that there are nevertheless grounds for regarding these competencies as distorting influences on our concepts. Our perfectly ordinary concept of causation (for instance), I suggest, is not normatively suffused. This is best brought out by thinking about science; I therefore begin with Knobe's claim that we ought not to understand folk judgments on the analogy of scientific hypothesis testing.
Knobe's claim that folk judgments are made in ways very unlike scientific hypothesis testing leaves us with a puzzle: Given that scientists are ordinary people too, how do they manage to engage in scientific research? If the relevant competencies are suffused with normative influences, how do scientists manage to free themselves of these influences (sufficiently well that they can identify them in the first place)? This question is important for several reasons, including, that if we can identify the means whereby scientists succeed in separating normative influences from the relevant judgments, we might all be in a position to make better normative judgments. At least on standard normative theories, our normative claims ought to follow from, rather than themselves cause, judgments of causation and intention; hence, separating out the normative from the non-normative might be a precondition of justified normative judgment.
So how do scientists manage to transcend the normative influences Knobe identifies? The answer is multifaceted, but an important part of it refers to the structure of the scientific enterprise. Science is an essentially distributed enterprise. The structure of a scientific community enables its members to compensate for the limitations and biases of individuals (Kitcher Reference Kitcher1993). Individual biases can thereby be cancelled out; one scientist's bias toward a hypothesis will be cancelled out by another's against it. Of course, this cancellation process is powerless against the kind of normative influences Knobe identifies, as they are universal.
But the structure of science has a second property: it externalizes scientific knowledge. Since science, by virtue of its essentially distributed nature, requires that data and theories be available to a multiplicity of researchers, they must be presented in a format that makes this possible, and that requires externalization. Once theories and data are externalized in this way, they become available for manipulation using formal techniques, and these techniques are designed to be impervious to the normative influences Knobe identifies. They can also be manipulated through the use of methods such as double blinding, which can also serve to filter out normative influences.
One implication of the forgoing is that the finding that ordinary people are not best understood on the model of scientists is unsurprising: no one is a scientist alone. An agent can be a scientist only as part of a community of researchers engaged in systematic inquiry. The contrast between scientific judgments and folk judgments is therefore misplaced: The contrast is not between different modes of thinking so much as between different ways of manipulating mental representations; one individualistic and the other deeply social.
An important implication is that there are grounds for seeing the competencies that agents utilize in making judgments as distortions of their concepts. We do not wish to say that scientists are mistaken in making causal judgments that are not normatively suffused. We therefore should not see the concept of causation as constituted by the structure of the competencies Knobe has elegantly uncovered. Scientists are members of the folk, and their onboard competencies are identical to everyone else's, yet they understand their causal judgments, qua scientists, as deploying the ordinary concept of causation, not a theoretical innovation. I suspect that given the choice between the concept of causation used in science and one that is explicitly normative, ordinary people would also choose the former, providing further evidence that scientists use the ordinary concept.
In saying this, I take issue neither with Knobe's arguments in favor of the view that our competencies are themselves normatively suffused, nor with his correlative claim that the rival view (according to which moral judgments bias our application of our concepts) is false. I am accepting that normative influences figure into the relevant competencies, but I am claiming that nevertheless we need to distinguish between these competences and the relevant concepts, even though we probably derive the concept from the competency (via some process of idealization). The concept of causation is normatively neutral, even though ordinary people deploy the concept using competencies that are normatively suffused.
It may be that we can dissociate the normatively neutral concept from the normatively suffused competencies only by externalizing and distributing our application of our concepts. We can hope to deploy our concepts better by becoming more like scientists. Doing so does not involve changing our onboard competencies – that may be a task that is beyond us – but instead requires that we alter the context in which we deploy them. By dividing and distributing cognitive labor, and by designing institutions that filter out the normative influences, we may become better reasoners, both in the normative and the non-normativerealms.
You have
Access
Joshua Knobe has added considerably to our knowledge of the ways in which ordinary people attribute intentions and make judgments regarding causation. In this commentary, I do not want to criticize his claim that the competencies agents deploy in making these judgments are deeply suffused with normative influences. However, I will suggest that there are nevertheless grounds for regarding these competencies as distorting influences on our concepts. Our perfectly ordinary concept of causation (for instance), I suggest, is not normatively suffused. This is best brought out by thinking about science; I therefore begin with Knobe's claim that we ought not to understand folk judgments on the analogy of scientific hypothesis testing.
Knobe's claim that folk judgments are made in ways very unlike scientific hypothesis testing leaves us with a puzzle: Given that scientists are ordinary people too, how do they manage to engage in scientific research? If the relevant competencies are suffused with normative influences, how do scientists manage to free themselves of these influences (sufficiently well that they can identify them in the first place)? This question is important for several reasons, including, that if we can identify the means whereby scientists succeed in separating normative influences from the relevant judgments, we might all be in a position to make better normative judgments. At least on standard normative theories, our normative claims ought to follow from, rather than themselves cause, judgments of causation and intention; hence, separating out the normative from the non-normative might be a precondition of justified normative judgment.
So how do scientists manage to transcend the normative influences Knobe identifies? The answer is multifaceted, but an important part of it refers to the structure of the scientific enterprise. Science is an essentially distributed enterprise. The structure of a scientific community enables its members to compensate for the limitations and biases of individuals (Kitcher Reference Kitcher1993). Individual biases can thereby be cancelled out; one scientist's bias toward a hypothesis will be cancelled out by another's against it. Of course, this cancellation process is powerless against the kind of normative influences Knobe identifies, as they are universal.
But the structure of science has a second property: it externalizes scientific knowledge. Since science, by virtue of its essentially distributed nature, requires that data and theories be available to a multiplicity of researchers, they must be presented in a format that makes this possible, and that requires externalization. Once theories and data are externalized in this way, they become available for manipulation using formal techniques, and these techniques are designed to be impervious to the normative influences Knobe identifies. They can also be manipulated through the use of methods such as double blinding, which can also serve to filter out normative influences.
One implication of the forgoing is that the finding that ordinary people are not best understood on the model of scientists is unsurprising: no one is a scientist alone. An agent can be a scientist only as part of a community of researchers engaged in systematic inquiry. The contrast between scientific judgments and folk judgments is therefore misplaced: The contrast is not between different modes of thinking so much as between different ways of manipulating mental representations; one individualistic and the other deeply social.
An important implication is that there are grounds for seeing the competencies that agents utilize in making judgments as distortions of their concepts. We do not wish to say that scientists are mistaken in making causal judgments that are not normatively suffused. We therefore should not see the concept of causation as constituted by the structure of the competencies Knobe has elegantly uncovered. Scientists are members of the folk, and their onboard competencies are identical to everyone else's, yet they understand their causal judgments, qua scientists, as deploying the ordinary concept of causation, not a theoretical innovation. I suspect that given the choice between the concept of causation used in science and one that is explicitly normative, ordinary people would also choose the former, providing further evidence that scientists use the ordinary concept.
In saying this, I take issue neither with Knobe's arguments in favor of the view that our competencies are themselves normatively suffused, nor with his correlative claim that the rival view (according to which moral judgments bias our application of our concepts) is false. I am accepting that normative influences figure into the relevant competencies, but I am claiming that nevertheless we need to distinguish between these competences and the relevant concepts, even though we probably derive the concept from the competency (via some process of idealization). The concept of causation is normatively neutral, even though ordinary people deploy the concept using competencies that are normatively suffused.
It may be that we can dissociate the normatively neutral concept from the normatively suffused competencies only by externalizing and distributing our application of our concepts. We can hope to deploy our concepts better by becoming more like scientists. Doing so does not involve changing our onboard competencies – that may be a task that is beyond us – but instead requires that we alter the context in which we deploy them. By dividing and distributing cognitive labor, and by designing institutions that filter out the normative influences, we may become better reasoners, both in the normative and the non-normativerealms.