Aquatic vertebrates evolved roughly 520 million years ago. The precise phylogenetic relationships connecting primitive jawless fish (ancestors of modern hagfish and lampreys) to modern vertebrates remains a matter of scientific discussion. Feinberg and Mallatt (Reference Feinberg and Mallatt2016) suggest that the ancestors of the earliest fish were worm-like organisms that lived in the thick microbial mat covering shallow sea floors in Ediacaran times. Once this resource started to diminish, some invertebrates moved out of the sludge and started living off other food sources, including each other, kicking off the Cambrian explosion of invertebrate forms. The ancestors of fish were slower to abandon the sludge but eventually they were forced to by the shrinking pools of microbial matter. They faced the challenge of locating and moving to the next pool without getting eaten by heavily armed arthropods.
This, according to Feinberg and Mallatt, drove a telemetry race between arthropod predators and their protovertebrate prey. The emergence of more advanced visual capabilities and integrated multisensory processing led, on their view, to conscious awareness of unified objects existing in the world. They locate the primary locus for this capacity in the optic tectum, which despite its name receives input from multiple senses. Feinberg and Mallatt surmise that its layered architecture allows common structure to be mapped among senses and organized into “a unified and coherent map of sensory space” (Feinberg & Mallatt, Reference Feinberg and Mallatt2016, p.108). The tectum also controls active, goal-driven motor behavior and has mechanisms for focusing attention. Thus, they argue, conscious perception of unified external objects likely arose in vertebrates over 500 million years ago.
Insofar as a case against this timeline can be inferred from Fleming and Michel’s target article, it is based on the idea they attribute to MacIver concerning the 100-fold increase in visual range that comes with living in air. Because the underwater visual horizon is comparatively very limited, especially in murky water, Fleming and Michel assert that to escape predation fish are left with “the only option for escape being a rapid and reflexive defensive maneuver” (p. 6).
These claims are overblown for several reasons.
While many ocean environments are murky, especially near river estuaries on rainy, temperate coasts, not all of them are. Both nutrient-poor tropical locales and nutrient-rich arctic habitats afford enough visibility to allows predators to be monitored (or prey stalked) from well beyond the range of the very rapid escape reflex due to Mauthner cells in the fish brain. (This reflex triggers in 5–10 ms, enabling a fish to turn and move as much as 1.5 body lengths inside 100 ms.) If you have spent a considerable amount of time trying to get close to fish in the ocean, which I have as a former spear-fisher and SCUBA instructor, you will know that many species of fish monitor your presence from a safe distance that gives them well more than 450 ms to respond to predator moves. In my experience, predators are detected early and carefully monitored. So, while the visual horizon underwater is much less than on land, it is not so reduced that it wouldn’t pay to build a world model that includes such threats.
We should always be cautious when generalizing about “fish.” Over 33,000 incredibly diverse species are known, comprising around 60% of all vertebrate species (Allen, Reference Allen2013). Fleming and Michel focus almost entirely on the responses by fish as prey, and even if this is how fish began their telemetry race in the Cambrian, they have evolved a huge range of predatory behavior, much of which depends on object tracking over extended periods of time. For instance, groupers and morays cooperatively pursue a single prey from one hiding place to another over many minutes (Bshary et al., Reference Bshary, Hohner, Ait-el-Djoudi and Fricke2006). While these species have evolved comparatively recently, their behaviors show that an ocean environment is not in itself an impediment to visually mediated, complex, non-reflexive behavior. Similarly, ocean-dwelling species have the capacity for trace conditioning over extended intervals (Nilsson et al., Reference Nilsson, Kristiansen, Fosseidengen, Ferno and van den Bos2008, Reference Nilsson, Kristiansen, Fosseidengen, Stien, Ferno and van den Bos2010). Trace conditioning is often cited as evidence for conscious processing because of the need to buffer a memory of the CS until the US arrives. Fleming and Michel dismiss trace conditioning as evidence for conscious processing on the grounds that decerebrate guinea pigs can be trace conditioned (ms. p.17, citing Kotani, Kawahara, & Kirino, Reference Kotani, Kawahara and Kirino2003). A careful reading of their source indicates that beyond 250 ms, cortical areas are required for stabilizing the association. The trace conditioning seen in the fish by Nilsson and colleagues is stable despite trace durations of many seconds. So, even if trace conditioning over very short intervals need not involve consciousness in decerebrate guinea pigs, it does not have the same signature as trace conditioning in fish.
Finally, Fleming and Michel’s arguments reveal an anthropocentric bias. Humans may not be a good yardstick for measuring the speed of conscious visual processing. It is well known, for instance, that many bird species have visual flicker-fusion rates more than double that of humans. Animals that can build world models faster than humans may find the capacity to do so useful even when visibility is reduced. Furthermore, some species of fish have adaptations to the relative lack of penetrability of water by light. Anchovies exploit the polarization of light to double the distance at which they can see prey (Novales Flamarique, Reference Novales Flamarique2019). Human blindness to polarization should not make us forget that there are many tricks available to biological evolution, even under the sea.
While the transition to land was no doubt momentous, it does not justify the whiggish account Fleming and Michel provide. Fish evolved for more than 100 million years before any of them took to the land, and fish have continued to evolve for another 400 million years since, expanding their visual and temporal horizons despite being stuck underwater. Progress toward understanding the origins and distribution of consciousness will require something better than single-factor just-so explanations of how our own consciousness came to be the way it is.
You have
Access
Aquatic vertebrates evolved roughly 520 million years ago. The precise phylogenetic relationships connecting primitive jawless fish (ancestors of modern hagfish and lampreys) to modern vertebrates remains a matter of scientific discussion. Feinberg and Mallatt (Reference Feinberg and Mallatt2016) suggest that the ancestors of the earliest fish were worm-like organisms that lived in the thick microbial mat covering shallow sea floors in Ediacaran times. Once this resource started to diminish, some invertebrates moved out of the sludge and started living off other food sources, including each other, kicking off the Cambrian explosion of invertebrate forms. The ancestors of fish were slower to abandon the sludge but eventually they were forced to by the shrinking pools of microbial matter. They faced the challenge of locating and moving to the next pool without getting eaten by heavily armed arthropods.
This, according to Feinberg and Mallatt, drove a telemetry race between arthropod predators and their protovertebrate prey. The emergence of more advanced visual capabilities and integrated multisensory processing led, on their view, to conscious awareness of unified objects existing in the world. They locate the primary locus for this capacity in the optic tectum, which despite its name receives input from multiple senses. Feinberg and Mallatt surmise that its layered architecture allows common structure to be mapped among senses and organized into “a unified and coherent map of sensory space” (Feinberg & Mallatt, Reference Feinberg and Mallatt2016, p.108). The tectum also controls active, goal-driven motor behavior and has mechanisms for focusing attention. Thus, they argue, conscious perception of unified external objects likely arose in vertebrates over 500 million years ago.
Insofar as a case against this timeline can be inferred from Fleming and Michel’s target article, it is based on the idea they attribute to MacIver concerning the 100-fold increase in visual range that comes with living in air. Because the underwater visual horizon is comparatively very limited, especially in murky water, Fleming and Michel assert that to escape predation fish are left with “the only option for escape being a rapid and reflexive defensive maneuver” (p. 6).
These claims are overblown for several reasons.
While many ocean environments are murky, especially near river estuaries on rainy, temperate coasts, not all of them are. Both nutrient-poor tropical locales and nutrient-rich arctic habitats afford enough visibility to allows predators to be monitored (or prey stalked) from well beyond the range of the very rapid escape reflex due to Mauthner cells in the fish brain. (This reflex triggers in 5–10 ms, enabling a fish to turn and move as much as 1.5 body lengths inside 100 ms.) If you have spent a considerable amount of time trying to get close to fish in the ocean, which I have as a former spear-fisher and SCUBA instructor, you will know that many species of fish monitor your presence from a safe distance that gives them well more than 450 ms to respond to predator moves. In my experience, predators are detected early and carefully monitored. So, while the visual horizon underwater is much less than on land, it is not so reduced that it wouldn’t pay to build a world model that includes such threats.
We should always be cautious when generalizing about “fish.” Over 33,000 incredibly diverse species are known, comprising around 60% of all vertebrate species (Allen, Reference Allen2013). Fleming and Michel focus almost entirely on the responses by fish as prey, and even if this is how fish began their telemetry race in the Cambrian, they have evolved a huge range of predatory behavior, much of which depends on object tracking over extended periods of time. For instance, groupers and morays cooperatively pursue a single prey from one hiding place to another over many minutes (Bshary et al., Reference Bshary, Hohner, Ait-el-Djoudi and Fricke2006). While these species have evolved comparatively recently, their behaviors show that an ocean environment is not in itself an impediment to visually mediated, complex, non-reflexive behavior. Similarly, ocean-dwelling species have the capacity for trace conditioning over extended intervals (Nilsson et al., Reference Nilsson, Kristiansen, Fosseidengen, Ferno and van den Bos2008, Reference Nilsson, Kristiansen, Fosseidengen, Stien, Ferno and van den Bos2010). Trace conditioning is often cited as evidence for conscious processing because of the need to buffer a memory of the CS until the US arrives. Fleming and Michel dismiss trace conditioning as evidence for conscious processing on the grounds that decerebrate guinea pigs can be trace conditioned (ms. p.17, citing Kotani, Kawahara, & Kirino, Reference Kotani, Kawahara and Kirino2003). A careful reading of their source indicates that beyond 250 ms, cortical areas are required for stabilizing the association. The trace conditioning seen in the fish by Nilsson and colleagues is stable despite trace durations of many seconds. So, even if trace conditioning over very short intervals need not involve consciousness in decerebrate guinea pigs, it does not have the same signature as trace conditioning in fish.
Finally, Fleming and Michel’s arguments reveal an anthropocentric bias. Humans may not be a good yardstick for measuring the speed of conscious visual processing. It is well known, for instance, that many bird species have visual flicker-fusion rates more than double that of humans. Animals that can build world models faster than humans may find the capacity to do so useful even when visibility is reduced. Furthermore, some species of fish have adaptations to the relative lack of penetrability of water by light. Anchovies exploit the polarization of light to double the distance at which they can see prey (Novales Flamarique, Reference Novales Flamarique2019). Human blindness to polarization should not make us forget that there are many tricks available to biological evolution, even under the sea.
While the transition to land was no doubt momentous, it does not justify the whiggish account Fleming and Michel provide. Fish evolved for more than 100 million years before any of them took to the land, and fish have continued to evolve for another 400 million years since, expanding their visual and temporal horizons despite being stuck underwater. Progress toward understanding the origins and distribution of consciousness will require something better than single-factor just-so explanations of how our own consciousness came to be the way it is.
Acknowledgments
The author would like to thank Adina Roskies for helpful comments and for encouraging him to be the kind of philosopher who thinks it is important to know a lot about fish.
Financial support
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.
Competing interests
None.