The TA reveals in excellent detail many species’ convergent evolution of cognition, one of the two great pillars of response selection. It seems unlikely that cognition could have developed before valence, the other pillar, since it would have been of no adaptive value without a response factor. “Valence” was introduced by Kurt Lewin to encompass reward and aversiveness together, as the force that moves individual choice: “Valence arguably is the fulcrum around which the dance of life revolves” (Lyon & Kuchling, Reference Lyon and Kuchling2021, 11). However, research on the basic mechanisms of valence has trailed behind research on cognition.

The authors list only three kinds of learning among their 34 cognitive functions (Target article, Figure 3): but valence gives adaptive value to all of those by inducing choice. Existing observations already suggest beginnings of cross-phyla research. The simple case of response shaping has been well studied with vertebrate subjects: Valence governs competition in a marketplace (Montague & Berns, 2002; Vanderveldt et al., 2016). Among nonhuman vertebrates, immature humans, and adult humans, when making spontaneous choices, a wide variety of rewards turn out to be commensurable with each other. So are aversive outcomes, though they are not the simple opposites of rewards (Pietri, Fazio, & Shook, Reference Pietri, Fazio and Shook2013). Recently, some invertebrates have also been reported to change their behaviors with experience, among them honeybees (Brill Meyer, & Rössler, Reference Brill, Meyer and Rössler2015; Leadbeater & Dawson, Reference Leadbeater and Dawson2017), jumping spiders (Edwards & Jackson, Reference Edwards and Jackson1994), and octopuses (Mather & Dickel, Reference Mather and Dickel2017; Richter, Hochner, & Kuba, Reference Richter, Hochner and Kuba2016). Evidence that the underlying neural mechanisms use the same, or analogous, neurotransmitters as vertebrates (dopamine and octopamine – Schwaerzel et.al., Reference Schwaerzel, Monastirioti, Scholz, Friggi-Grelin, Birman and Heisenberg2003) suggests that these mechanisms may share the same basic properties: dependence on appetite, variation with fatigue, diet, and environment, and, of greatest interest for the study of human evolution, attenuation with delay from the moment of choice.

Attenuation by delay has been widely demonstrated with vertebrates, usually in a hyperboloid curve (Vanderveldt et al., 2016), even when an effect of reward immediacy per se can be ruled out (Green, Myerson, & Macaux, Reference Green, Myerson and Macaux2005; Salas-Morellón, Palacios-Huerta, & Call, Reference Salas-Morellón, Palacios-Huerta and Call2024). The effect of delay on valence could readily be visualized with invertebrates – for nonexponential discounting, picture an octopus going to a closer hole to pull through smaller, sooner (SS) reward, but heading to a farther hole to pull through a larger, later (LL) reward when the holes are at greater distances – but with the same separation (as a follow-up of Mather & Dickel, Reference Mather and Dickel2017). The additive property of reward in multiple choices has been shown in rats (Ainslie & Monterosso, Reference Ainslie and Monterosso2003) and could be tested similarly with invertebrates. Withholding SS consumption has been observed in many nonhuman vertebrate species (Brucks et.al., Reference Brucks, Petelle, Baldoni, Krasheninnikova, Rovegno and von Bayern2022; Miller et.al., Reference Miller, Boeckle, Jelbert, Frohnwieser, Wascher and Clayton2019), as has prospection (“time travel” – Mulcahy & Call, Reference Mulcahy and Call2006; Osvath & Martin-Ordas, Reference Osvath and Martin-Ordas2014; Raby et.al., Reference Raby, Alexis, Clayton and Dickinson2007). Presumably, the latter feats would be a stretch for invertebrates, but could be tested for.

Like convergent evolutions of cognition, convergent evolution of valence might identify basic mechanisms that shed light on some persistent theoretical challenges. Among these:

• The emergence of opioid “hot spots” dotting various parts of the vertebrate brain (Berridge, Reference Berridge2018) suggests the evolutionary superimposition of relatively long-term “liking” valence to compensate for the short-term focus of the ostensibly well-adapted “wanting” system. These two systems have already been studied in insects (Perry & Barron, Reference Perry and Barron2013).

• The combination of attracting attention but deterring motor behavior that characterizes negative “incentive salience” might be more separable in simpler nervous systems. In vertebrates, neurons modulating each inhabit the same neighborhood, anatomically and functionally – contextual cues switch the very same dopaminergic neurons between reward and aversiveness (Berridge, Reference Berridge2018).

• Classical conditioning has usually been thought to occur without regard to valence, rather as automatic learning by association with unconditioned responses. However, stimuli without detectable valence cannot induce conditioning (Miller, Reference Miller1969), leading some authors to suggest that conditioning is a form of operant learning (Donahoe, Reference Donahoe2021; Donahoe, Burgos, & Palmer, Reference Donahoe, Burgos and Palmer1993; Pear & Eldridge, Reference Pear and Eldridge1984). Reward with a hyperboloid time course could lure agents into entrapment by aversive experiences (Ainslie, Reference Ainslie2005, 637–639), with such a lure experienced as an urge – for instance, to panic, to despair, or to rage. Cross-phyla research might suggest how such a time-sensitive mechanism evolved

  There is the further likelihood that many unconditioned responses which have been assumed to be reflexive – hardwired – turn out to be modifiable according to their results: softwired, if you will. That is, they are default patterns that are followed only in the absence of experience with focused contrary valence. For instance, bulimic patients say they have learned to vomit at will first by physical gagging, then by imagining it, then by simply intending it. Hatha yoga is said to teach voluntary control of many such “involuntary” behaviors (Wenger & Bagchi, Reference Wenger and Bagchi1961).

• Emotions are now widely recognized as having valence. Like the “conditionable” responses in the previous bullet, they may turn out to be chosen on the basis of this valence, rather than being reflexively triggered and classically conditioned. This possibility, raised unequivocally by Sartre (Reference Sartre1939/Reference Sartre1948) and obliquely by others (e.g. “Emotion states exist somewhere between reflexes and deliberate action” – Lyon & Kuchling, Reference Lyon and Kuchling2021, 4), would mean that the time course of an emotion’s valence can be the basis, alternatively, of actively seeking it, or of avoiding it as a Siren, or of being torn by ambivalence.

• Furthest evolved from external contingency, intrinsic motivation is becoming recognized as highly plastic (Morris et al., Reference Morris, Grehl, Rutter, Mehta and Westwater2022). A tonic core affect has been suggested to be continuously operating in humans (Russell, Reference Russell2009; Lyon & Kuchlin, Reference Lyon and Kuchling2021), like the circulation of the blood. Learned manipulation of that affect may be what supplies motivation for fantasy, unrewarded games, and abstract cognition, without the need for eventual inbred rewards (Ainslie, Reference Ainslie2017, 178–183, and Reference Ainslie2023, 7–10). Conversely, loss of core affect may produce anhedonia (Der-Avakian & Markou, Reference Der-Avakian and Markou2012; Pizzagalli, Reference Pizzagalli2022). Experientially, habits are automatic – no longer motivated – but failure to perform a habitual action provokes discomfort, perhaps by interruption of tonic reward that was actively maintaining it.

Rudiments of such phenomena in less evolved organisms might suggest their mechanisms.