Cumulative culture – the social transmission and accumulation of innovations across generations – has enabled humans to develop technologies of astonishing complexity (Boyd & Richerson Reference Boyd and Richerson1996). Osiurak and Reynaud (O&R) present a timely re-evaluation of the potential cognitive capacities enabling this process, shifting the focus from social-learning to technical-reasoning skills. We enjoyed O&R's fresh perspective, but were surprised that they did not examine more carefully their starting assumption that only humans possess cumulative technological culture. Two non-human species have been suggested to refine their foraging tools in a cumulative manner – one of our closest relatives, the chimpanzee, and a tropical corvid, the New Caledonian crow (Dean et al. Reference Dean, Vale, Laland, Flynn and Kendal2014). Although O&R mention New Caledonian crows in passing, we feel they have missed a valuable opportunity to consult the large body of evidence available for this non-primate model system. Here, we provide some of this missing context, and argue that debates about cumulative technological culture benefit greatly from the comparative insights afforded by these remarkable birds.

According to a recent consensus definition (Mesoudi & Thornton Reference Mesoudi and Thornton2018), the New Caledonian crow is a compelling candidate for cumulative technological culture (Hunt & Uomini Reference Hunt and Uomini2016; St Clair et al. Reference St Clair, Klump, Sugasawa, Higgott, Colegrave and Rutz2018). A detailed evaluation is beyond the scope of this commentary and will be provided elsewhere, but the key points can be summarised as follows. First, New Caledonian crows' technology exhibits clear signs of cumulative refinement. In addition to using unmodified sticks and other objects, the species actively manufactures two elaborate tool types from living plant material: hooked stick tools and barbed leaf tools (Hunt Reference Hunt1996; Rutz & St Clair Reference Rutz and St Clair2012). For both of these, distinct variants have been described that strongly suggest past accumulation of design innovations, progressing from basic to increasingly complex tools (Hunt & Gray Reference Hunt and Gray2003; St Clair & Rutz Reference St Clair and Rutz2013). Second, controlled experiments have recently demonstrated that hooked stick tools are much better for extracting prey across a variety of foraging scenarios than (the presumably ancestral) non-hooked stick tools (St Clair et al. Reference St Clair, Klump, Sugasawa, Higgott, Colegrave and Rutz2018), and that well-crafted, deep hooks are particularly efficient (Sugasawa et al. Reference Sugasawa, Klump, St Clair and Rutz2017). Third, it seems likely that at least some tool-related information is passed across generations via social learning, leading to a rudimentary form of material culture, with striking regional variation in tool behaviour (Hunt & Gray Reference Hunt and Gray2003; St Clair et al. Reference St Clair, Klump, van der Wal, Sugasawa and Rutz2016; see also Bluff et al. Reference Bluff, Kacelnik and Rutz2010a). Here, the evidence is weaker (review: Rutz et al. Reference Rutz, Hunt and St Clair2018), but field studies have shown that juvenile crows hone their tool-related skills during an unusually long dependency period (Bluff et al. Reference Bluff, Troscianko, Weir, Kacelnik and Rutz2010b; Hunt et al. Reference Hunt, Holzhaider and Gray2012), which affords abundant opportunities to observe experienced adults (e.g., Holzhaider et al. Reference Holzhaider, Hunt and Gray2010a; Rutz et al. Reference Rutz, Burns, James, Ismar, Burt, Otis, Bowen and St Clair2012; St Clair et al. Reference St Clair, Burns, Bettaney, Morrissey, Otis, Ryder, Fleischer, James and Rutz2015), and laboratory experiments have confirmed a capacity to learn both from demonstrators (Kenward et al. Reference Kenward, Rutz, Weir and Kacelnik2006; Logan et al. Reference Logan, Breen, Taylor, Gray and Hoppitt2016) and artefacts (Jelbert et al. Reference Jelbert, Hosking, Taylor and Gray2018).

O&R identified technical-reasoning skills as a potential cognitive prerequisite of cumulative technological culture. New Caledonian crows' tool repertoire is obviously rudimentary compared to human technology, consisting entirely of one-piece artefacts crafted from plant materials, but the system allows us to chart the minimal cognitive capacities required for the accumulation of design improvements. Some 20 years of research on the species' tool-related cognition provide instructive insights. New Caledonian crows have been shown: to make and/or select task-appropriate tools (e.g., Bluff et al. Reference Bluff, Weir, Rutz, Wimpenny and Kacelnik2007; Reference Bluff, Troscianko, Weir, Kacelnik and Rutz2010b; Chappell & Kacelnik Reference Chappell and Kacelnik2002; Reference Chappell and Kacelnik2004; Knaebe et al. Reference Knaebe, Taylor, Elliffe and Gray2017), most likely relying on simple “heuristics” (Hunt et al. Reference Hunt, Rutledge and Gray2006; G.R. Hunt, unpublished data); to attend to different tool design features (Knaebe et al. Reference Knaebe, Taylor, Miller and Gray2015; St Clair & Rutz Reference St Clair and Rutz2013; but see Holzhaider et al. Reference Holzhaider, Hunt, Campbell and Gray2008); and to learn quickly – and even invent – novel tool behaviours (Rutz, Reference Rutzunpublished manuscript). Although these skills are impressive, reports of more advanced cognitive abilities have routinely been contested. This includes claims that New Caledonian crows are capable of: tool-related causal reasoning (pro: Taylor et al. Reference Taylor, Elliffe, Hunt, Emery, Clayton and Gray2011; contra: Ghirlanda & Lind Reference Ghirlanda and Lind2017; Hennefield et al. Reference Hennefield, Hwang, Weston and Povinelli2018); analogical reasoning in the context of sequential tool use (pro: Taylor et al. Reference Taylor, Hunt, Holzhaider and Gray2007; contra: Wimpenny et al. Reference Wimpenny, Weir, Clayton, Rutz and Kacelnik2009; and see: Taylor et al. Reference Taylor, Elliffe, Hunt and Gray2010); making inferences about hidden causal agents (pro: Taylor et al. Reference Taylor, Miller and Gray2012b; contra: Boogert et al. Reference Boogert, Arbilly, Muth and Seed2013; Dymond et al. Reference Dymond, Haselgrove and McGregor2013); creating novel causal interventions (pro: Jacobs et al. Reference Jacobs, von Bayern, Martin-Ordas, Rat-Fischer and Osvath2015; contra: Taylor et al. Reference Taylor, Cheke, Waismeyer, Meltzoff, Miller, Gopnik, Clayton and Gray2014; and see: von Bayern et al. Reference von Bayern, Heathcote, Rutz and Kacelnik2009); adaptively fashioning new tool shapes (pro: Weir et al. Reference Weir, Chappell and Kacelnik2002; contra: Rutz et al. Reference Rutz, Sugasawa, van der Wal, Klump and St Clair2016b); and using mental representations to solve tool problems (pro: Gruber et al. Reference Gruber, Schiestl, Boeckle, Frohnwieser, Miller, Gray, Clayton and Taylor2019; contra: Vonk Reference Vonk2019). These disagreements may partly reflect the methodological challenges inherent in probing the minds of non-verbal subjects, but in our view, they also bring into focus the upper limits of the species' reasoning abilities.

In conclusion, New Caledonian crows excel at solving tool-related tasks, exhibiting a clear capacity for quick learning and behavioural flexibility, but evidence of more advanced cognitive abilities is much weaker. This suggests that a rudimentary form of cumulative technological advancement is possible without some of the technical-reasoning skills O&R predicted to be essential (as well as without language and active teaching; see Bluff et al. Reference Bluff, Kacelnik and Rutz2010a; Holzhaider et al. Reference Holzhaider, Hunt and Gray2010a). If our preliminary assessment is correct, it would raise the possibility that cognitive constraints are currently limiting the scope for further technological advancement in New Caledonian crows, and/or that different mechanisms enable cumulative technology in humans and New Caledonian crows. Although non-cultural processes are expected to contribute to shaping the technology of New Caledonian crows (Kenward et al. Reference Kenward, Weir, Rutz and Kacelnik2005; Rutz et al. Reference Rutz, Hunt and St Clair2018), we should not underestimate the extent to which this was also the case in our tool-making ancestors (e.g., Corbey et al. Reference Corbey, Jagich, Vaesen and Collard2016). Therefore, whatever the precise contributions of genetics, individual learning and social transmission (Hunt & Uomini Reference Hunt and Uomini2016; Rutz et al. Reference Rutz, Hunt and St Clair2018), the New Caledonian crow – and other non-human tool users (e.g., Rutz et al. Reference Rutz, Klump, Komarczyk, Leighton, Kramer, Wischnewski, Sugasawa, Morrissey, James, St Clair, Switzer and Masuda2016a) – provide a powerful comparative framework (Dean et al. Reference Dean, Vale, Laland, Flynn and Kendal2014) for identifying the basic biological processes that enable cumulative technological evolution.