In What Babies Know, Spelke (Reference Spelke2022) uses the framework of early human development to uncover the perceptual and cognitive architecture of humanity. Not only is this a compelling argument. It is also very clearly the right position to take when confronted with the empirical evidence from across all domains of psychology. Her very first aim is to explore what infants know at the time when their learning begins. At this initial hurdle, What Babies Know immediately becomes contentious. Learning does not begin during infancy. There is now a wealth of data to show that learning across several cognitive domains is present during the third trimester of pregnancy. When viewed through the lens of the fetus, What Babies Know offers some intriguing possibilities for the future of research during this earlier period of development.

Thankfully, the fetus has not been entirely ignored in What Babies Know. When the topic of language processing is broached, seminal work such as DeCaspar and Fifer (Reference DeCaspar and Fifer1980) is acknowledged and at the forefront of the chapter. What is surprising is that given the aims of the book, the fetus is mystifyingly absent when an in-depth exploration of this topic would enrich the arguments present throughout. Given that Spelke is aware of the fetus, how does the fetus interplay with the thesis that “experiments on infants provide the most direct access to the earliest emerging cognitive capacities at the foundations of our knowledge?” (Spelke, Reference Spelke2022, prologue xxi). It is the fetus, not the infant, that aligns with this statement. Given the carefully constructed positions that are developed throughout the book, there must be a rationale for the exclusion of evidence in multiple areas, despite fetal data having direct relevance to several topics in the volume.

The third trimester is so important for the development of perception and cognition that it has clear implications for a number of sections in What Babies Know. For example, number processing has been explored via neuroscience methods in the fetus (Schleger et al., Reference Schleger, Landerl, Muenssinger, Draganova, Reinl, Kiefer-Schmidt and Preissl2014). Fetal vision is also advancing as a field, now that technical impediments are being overcome (e.g., Reid et al., Reference Reid, Dunn, Young, Amu, Donovan and Reissland2017). In the domain of object processing, eye movements related to tracking an object in space, in this case light on the maternal abdomen, have also been observed via ultrasound (Donovan, Dunn, Penman, Young, & Reid, Reference Donovan, Dunn, Penman, Young and Reid2020). In the area of vision and object processing, Spelke does outline some of the seminal work on the biological basis of fetal development (e.g., as reviewed in Ackman & Crair, Reference Ackman and Crair2014) and provides suggestions on why the fetus may have object processing capacities and emerging spatial function. Work with the fetus nonetheless shines a clearer light on the parameters of other domains covered in the book. With speech detection and early learning, for example, the melody of a neonate cry is shaped by auditory experiences prior to birth (Mampe, Friederici, Cristophe, & Wermke, Reference Mampe, Friederici, Cristophe and Wermke2009). There is a role for experience to shape the parameters of speech detection during the fetal period. There has been an exciting growth in our knowledge related to the perceptual and psychological characteristics of the fetus (Reid & Dunn, Reference Reid and Dunn2021), with recent models of sound processing (Vogelsang, Vogelsang, Diamond, & Sinah, Reference Vogelsang, Vogelsang, Diamond and Sinah2023) demonstrating the impact of auditory processing on the spoken word during pregnancy. Clearly the fetus has much to offer when considering the themes that are outlined by Spelke.

What Babies Know also maps the evolution of thinking surrounding core knowledge since initial proposals several decades ago through to the present day. Within this framework, the changing definitions of the term innate are of particular interest. The current position, namely “they are present and functional on an infant's first informative encounters with the entities that they serve to represent,” (Spelke, Reference Spelke2022, prologue xix) actively incorporates aspects of constructivist viewpoints. Through so doing, Spelke clearly seeks to remove fuel from the fire of the historically heated debate on the role of experience in shaping development (e.g., Spelke & Kinzler, Reference Spelke and Kinzler2009; Spencer et al., Reference Spencer, Samuelson, Blumberg, McMurray, Robinson and Tomblin2009).

If encountering information prior to when it is required is an essential aspect of cognitive development, then under this definition of innate, the nature of acquisition is incorporated into the framework of innate processes. With this premise, learning and the seeking out of specific forms of information to refine neural development are the natural unfolding of genetic codes. This approach effectively signals an acknowledgement that learning is inherent in human development and much of the neural systems that underpin cognitive development are not hard wired in genetics. Rather, it is the pathway to learning itself that is now defined in nativist terms. Such an approach is capable of encompassing everything from the formation of initial visual experience via retinal waves occurring before a fully functional visual system (e.g., Firth, Wang, & Feller, Reference Firth, Wang and Feller2005) which is beautifully described by Spelke, through to the proposals that teaching is an inherent part of being human (Csibra & Gergely, Reference Csibra and Gergely2011). There has been a significant accumulation of evidence in recent decades that supports a neuroconstructivist approach to development. For example, computational modeling on the development of processing the past tense (Westermann & Ruh, Reference Westermann and Ruh2012) presents a strong argument for this position. The approach by Karmiloff-Smith (Reference Karmiloff-Smith2009), when exploring developmental disorders is equally compelling. Spelke's modification to the innate position will undoubtedly produce much thought and dialogue in the field.

From the perspective of fetal research, it is important to recognize that the methods outlined by Spelke for application with infant populations are almost certainly those best suited for research with the fetus. Fetal research was once almost exclusively the domain of observation (e.g., Zoia et al., Reference Zoia, Blason, D'Ottavio, Bulgheroni, Pezzetta, Scabar and Castello2007) or correlations with aspects of fetal development with later behavioral outcomes, such as prenatal head circumference and later educational achievement (e.g., DiPietro, Costigan, & Voegtline, Reference DiPietro, Costigan and Voegtline2015). Due to the advent of noninvasive brain imaging and high-quality ultrasound, experimental approaches are increasingly common with the fetus. The paradigms utilized are exclusively derived from infancy research (e.g., Johnson & Morton, Reference Johnson and Morton1991) or from the neurosciences (e.g., oddball paradigms, for a review, see Dunn, Reissland, & Reid, Reference Dunn, Reissland and Reid2015). Fetal research is by no means a fully established field. There are currently no mapped parameters related to fetal attention to auditory, tactile, or visual stimulation. There is little point in utilizing a paradigm where your sample is unable to process the stimuli due to fatigue or inattention. The consequence of this is that current experiments are short in nature and typically feature only two conditions. What is also unknown are the specific infancy paradigms that are best suited to the fetal population. Spelke highlights successful and illuminating experiments with the infant that utilize peripheral stimulus presentation, paired preference, sequential presentations, and classic habituation amongst even more paradigms. The volume reads like a shopping list of experimental methods that are yet to be employed and fully understood with a fetal sample.

The arguments presented in What Babies Know raises a number of important questions related to the fetal period. Can motion be discriminated during the third trimester? If so, will there be preferential engagement with agentive motion as is the case with the newborn (e.g., Di Giorgio, Lunghi, Simion, & Vallortigara, 2017)? It is tempting to determine that fetal processing of information must derive from genetic predispositions. Such a conclusion does not account for fetal experiences that must be considered. For example, it is possible that self-produced actions (Myowa-Yamakoshi & Takeshita, Reference Myowa-Yamakoshi and Takeshita2006; Wilkinson, Paikan, Gredebäck, Rea, & Metta, Reference Wilkinson, Paikan, Gredebäck, Rea and Metta2014) and the visual capacity to monitor them (Del Guidice, 2011) could lay the foundation that drives early psychological systems, including a preference for agentive motion. It is also feasible that contingently responsive maternal behavior produces environmental change that is detected as causal by the fetus. Supporting this notion, prior work has shown that fetal heart rate is modified based on maternal swinging movements (Lecanuet & Jacquet, Reference Lecanuet and Jacquet2002). Based on these studies, an experimental framework for social-cognitive development is possible, even during the fetal period. From this specific topic alone, it is clear that in order to understand the issue of core knowledge, we must start not with the infant, but with the fetus. Spelke is unquestionably one of the most preeminent scientists of the past century. It is my hope that the equivalent to Spelke born in the 2000s will, after many decades of research, produce a volume entitled What the Fetus Knows.