Fleming and Michel (F&M) argue that many of our responses to the world can be mediated by unconscious processes, and they provide compelling evidence of fast unconscious processes able to mediate action in humans and throughout the evolutionary tree. If so much can be done unconsciously, what is conscious sensory processing for? And how does F&M’s discussion bear upon recent challenges to the efficacy of consciousness in directing human action?
According to F&M, consciousness is important for model-based planning, a need engendered by extended sensory horizons. Model-based planning involves (usually unconscious) simulations on a world model, and when a system encompasses both online perception and simulation, these two informational streams need to be distinguished. Thus, they suggest that consciousness is related to perceptual reality monitoring. As models of perceptual reality must be sensitive to estimates of reliability and should modulate evidence gathering, F&M argue that what we are conscious of is a constructed world-model that passes some threshold for reliability as indicated by subjective confidence.
I am generally in agreement with their account of the function of consciousness as identifying a high-credence model to support flexible, deliberative planning and decision-making. This provides a partial answer to one of the most vexing questions in both philosophy of mind and evolutionary theory. It is partial because nothing about their story (or any extant story) answers the deepest puzzle about the function of consciousness, which is why it has phenomenal character. In their view, consciousness is meant to mark the reliable and fixed version of a world model, but information can be marked as special in many ways. Their story leaves unaddressed why (and how) its marking should result in phenomenality.
Despite finding their functional story plausible, the way in which they tell this story may lead to some misconceptions regarding how and when consciousness functions in ordinary agency. What follows is more a clarification than a criticism of F&M’s view.
F&M suggest that consciousness is slow: on the basis of postdictive effects on conscious reports, they estimate that it takes approximately 350–400 ms for a visual stimulus to reach conscious awareness. The implication is that our awareness lags significantly behind the worldly present and that, for the most part, our reactions to the world occur unconsciously. This claim needs to be scrutinized in order to avoid misapprehensions about what the temporal profile of consciousness means for human agency and for free will.
It is undeniably true that neural processing takes time, and thus, there are limits to what we can respond to, unconsciously or consciously. Even so, the timing data for conscious experience seems surprisingly slow. F&M canvas the literature on postdictive visual illusions, trying to fix on a number to characterize the time of consciousness, which they identify as approximately 350–400 ms. Their discussion implies that there is a correct single number to be found, but another possibility exists: that different aspects of perception are processed with different time courses. Such a phenomenon is commonplace, as when one becomes aware of having touched something scalding before registering the pain the burn brings. With respect to the Vernier effect F&M describe and other visual illusions, it is possible that the fine details of a visual stimulus take longer to reach consciousness than more general characteristics, and the time-course of retrodictive effects may depend on reliability, magnitude, and so on. If consciousness is related to confidence, as they suggest, then it is likely that different aspects of the world are available to consciousness at different times. Global masking effects that prevent any conscious percept may suggest otherwise, but they are also difficult to interpret because of potential interference with memory and response processing.
The point remains that dynamic aspects of consciousness should not be ignored or underestimated. These dynamic aspects may also be related to the temporal profile of what is experienced. As William James recognized more than a century ago, the experienced present is not an instant, but a temporal slice of some thickness that already encodes information about the future just as it does about the past. This is the “specious present,” a construct that fits well with F&M’s view that the contents of consciousness are predictive models.
One reason all this matters is that it bears upon deep questions about human agency. For example, many people already erroneously assent to the claim that most of our actions are initiated unconsciously and that “our brains decide before we do,” largely based on a common misunderstanding of experimental results involving the “readiness potential” from Benjamin Libet dating from the 1980s (Libet et al, Reference Libet, Gleason, Wright and Pearl1983; Libet Reference Libet1985). This view has negatively influenced discussions of free will for the past forty years, despite the fact that the data do not warrant this common interpretation (see, e.g., Schurger et al, Reference Schurger, Hu, Pak and Roskies2021).
The discussion of F&M about the speed of consciousness is likely to feed into the very same misconceptions, for it implies that our conscious selves are always half a second behind reality. If this were correct, we would be in much more precarious situations than we think we are both practically and morally, when, for instance, driving at high speeds or conversing with others.
However, when properly understood in the context of the predictive coding model that F&M champions, lags in consciousness of external stimuli are consistent with our ongoing conscious involvement with the world in ways that preserve most intuitive conceptions of agency, free will, and responsibility. If we are indeed constructing conscious models of the world around us that incorporate predictive information (including being responsive to violations of predictions), then we are usually conscious of what is going on around us as it is unfolding, and we are consciously responding to that information in all but the most unusual circumstances. The existence of postdictive effects on verbal reports doesn’t imply that we are only ever conscious of the world with a significant lag. Nor should the fact that consciousness is a process that takes time be taken as support for a neuroscientific argument against free will.
You have
Access
Fleming and Michel (F&M) argue that many of our responses to the world can be mediated by unconscious processes, and they provide compelling evidence of fast unconscious processes able to mediate action in humans and throughout the evolutionary tree. If so much can be done unconsciously, what is conscious sensory processing for? And how does F&M’s discussion bear upon recent challenges to the efficacy of consciousness in directing human action?
According to F&M, consciousness is important for model-based planning, a need engendered by extended sensory horizons. Model-based planning involves (usually unconscious) simulations on a world model, and when a system encompasses both online perception and simulation, these two informational streams need to be distinguished. Thus, they suggest that consciousness is related to perceptual reality monitoring. As models of perceptual reality must be sensitive to estimates of reliability and should modulate evidence gathering, F&M argue that what we are conscious of is a constructed world-model that passes some threshold for reliability as indicated by subjective confidence.
I am generally in agreement with their account of the function of consciousness as identifying a high-credence model to support flexible, deliberative planning and decision-making. This provides a partial answer to one of the most vexing questions in both philosophy of mind and evolutionary theory. It is partial because nothing about their story (or any extant story) answers the deepest puzzle about the function of consciousness, which is why it has phenomenal character. In their view, consciousness is meant to mark the reliable and fixed version of a world model, but information can be marked as special in many ways. Their story leaves unaddressed why (and how) its marking should result in phenomenality.
Despite finding their functional story plausible, the way in which they tell this story may lead to some misconceptions regarding how and when consciousness functions in ordinary agency. What follows is more a clarification than a criticism of F&M’s view.
F&M suggest that consciousness is slow: on the basis of postdictive effects on conscious reports, they estimate that it takes approximately 350–400 ms for a visual stimulus to reach conscious awareness. The implication is that our awareness lags significantly behind the worldly present and that, for the most part, our reactions to the world occur unconsciously. This claim needs to be scrutinized in order to avoid misapprehensions about what the temporal profile of consciousness means for human agency and for free will.
It is undeniably true that neural processing takes time, and thus, there are limits to what we can respond to, unconsciously or consciously. Even so, the timing data for conscious experience seems surprisingly slow. F&M canvas the literature on postdictive visual illusions, trying to fix on a number to characterize the time of consciousness, which they identify as approximately 350–400 ms. Their discussion implies that there is a correct single number to be found, but another possibility exists: that different aspects of perception are processed with different time courses. Such a phenomenon is commonplace, as when one becomes aware of having touched something scalding before registering the pain the burn brings. With respect to the Vernier effect F&M describe and other visual illusions, it is possible that the fine details of a visual stimulus take longer to reach consciousness than more general characteristics, and the time-course of retrodictive effects may depend on reliability, magnitude, and so on. If consciousness is related to confidence, as they suggest, then it is likely that different aspects of the world are available to consciousness at different times. Global masking effects that prevent any conscious percept may suggest otherwise, but they are also difficult to interpret because of potential interference with memory and response processing.
The point remains that dynamic aspects of consciousness should not be ignored or underestimated. These dynamic aspects may also be related to the temporal profile of what is experienced. As William James recognized more than a century ago, the experienced present is not an instant, but a temporal slice of some thickness that already encodes information about the future just as it does about the past. This is the “specious present,” a construct that fits well with F&M’s view that the contents of consciousness are predictive models.
One reason all this matters is that it bears upon deep questions about human agency. For example, many people already erroneously assent to the claim that most of our actions are initiated unconsciously and that “our brains decide before we do,” largely based on a common misunderstanding of experimental results involving the “readiness potential” from Benjamin Libet dating from the 1980s (Libet et al, Reference Libet, Gleason, Wright and Pearl1983; Libet Reference Libet1985). This view has negatively influenced discussions of free will for the past forty years, despite the fact that the data do not warrant this common interpretation (see, e.g., Schurger et al, Reference Schurger, Hu, Pak and Roskies2021).
The discussion of F&M about the speed of consciousness is likely to feed into the very same misconceptions, for it implies that our conscious selves are always half a second behind reality. If this were correct, we would be in much more precarious situations than we think we are both practically and morally, when, for instance, driving at high speeds or conversing with others.
However, when properly understood in the context of the predictive coding model that F&M champions, lags in consciousness of external stimuli are consistent with our ongoing conscious involvement with the world in ways that preserve most intuitive conceptions of agency, free will, and responsibility. If we are indeed constructing conscious models of the world around us that incorporate predictive information (including being responsive to violations of predictions), then we are usually conscious of what is going on around us as it is unfolding, and we are consciously responding to that information in all but the most unusual circumstances. The existence of postdictive effects on verbal reports doesn’t imply that we are only ever conscious of the world with a significant lag. Nor should the fact that consciousness is a process that takes time be taken as support for a neuroscientific argument against free will.
Acknowledgements
None.
Financial support
The author has no funding sources to declare.
Competing interests
The author has no competing interests.