Hazardous tools: the emergence of reasoning in human tool use

For the majority of the manipulation-based theorists, when we see a tool, we automatically activate two kinds of motor representations: How to grasp it (i.e., “structural affordances”) and How to use it (i.e., “functional affordances”; Buxbaum & Kalénine, 2010; Thill et al., 2013; Bach, Nicholson, & Hudson, 2014; Kourtis & Vingerhoets, 2015; Kourtis, Vandemaele, & Vingerhoets, 2018). By positing the automatic activation of motor representations associated with the tool’s usual function (i.e., How to use it), the so-called automatic activation hypothesis of functional affordances becomes a crucial aspect of many embodied approaches to human tool use (e.g., Bach et al., 2014; Buxbaum, 2001; Gonzalez Rothi, Ochipa, & Heilma, 1991; van Elk, van Schie, & Bekkering, 2014). Please note that the terms “gesture engram” (Buxbaum, 2001), “visuokinesthetic motor engram” (Heilman, Gonzalez Rothi, & Valenstein, 1982), “spatial–temporal movement representation” (Heilman & Watson, 2008), “manipulation knowledge” (Bach et al., 2014; van Elk et al., 2014) or “representation of motor programs for acquired tool use skills” (Johnson-Frey, Newman-Norlund, & Grafto, 2005), as well as many others, are generally used as synonyms of functional affordances. Thus, while it differs from the initial conceptualisation made by Gibson (1977) and despite its polysemic nature, the concept of functional affordance remains univocally meaningful.

With the term “mechanical knowledge” (within this paper, synonym of “technical reasoning”) we considered a form of knowledge about physical principles. Such a non-declarative kind of knowledge may be considered as abstract because physical and technical realities do not overlap. Hence, a single physical matter (e.g., glass) can have distinct properties (e.g., hardness, sharpness, transparency). Conversely, distinct physical matters (e.g., metal or plastic) can have the same single property (e.g., hardness). Additionally, as happens in by-analogy problem solving situations, we are able to quickly transfer the mechanical principles we learned through a specific tool or in a specific circumstance to another (e.g., a knife may be used to screw by transferring the function usually associated with a screwdriver to the knife). Crucially, one of the peculiarities of human tool use lies exactly in transfer skills (e.g., Penn, Holyoak, & Povinelli, 2008). Further discussions on these aspects may be found in Osiurak and Badets (2016).

Given the background of the study, one might wonder why we preferred to study oculomotor behaviour rather than actual tool manipulation behaviour (e.g., motor preparation of grasps or other typical affordance-related experimental tasks). Alongside the relevant eye-hand coordination issues we introduced above, here it may be useful to remind that the basic assumption of the direct-visual-route-to-action view is that vision guides action (Milner & Goodale, 2008). In particular, it has been repeatedly suggested in the literature how the “implicit recognition of action-related object attributes [affordances] can bias object competition—and visual spatial attention” so that “the motor affordance of an object must first be recognized, a process that likely involves attention to specific object features” (Handy, Grafton, Shroff, Ketay, & Gazzaniga, 2003, p. 424–425). Therefore, the temporal allocation of visuo-spatial attention reflects visually guided behaviour. If an affordance associated with a tool is perceived, the fixation pattern of the tool will be mostly focused on its action-relevant parts (Handy et al., 2003; Land, 2006; Roberts & Humphreys, 2011; Natraj, Pella, Borghi, & Wheaton, 2015; Ambrosini & Costantini, 2017). Consistently with that, the temporal allocation of visuo-spatial attention to specific areas of tools (e.g., the handle of a hammer) can be effectively used as an indirect index of affordance perception (see Federico & Brandimonte, 2019 for a detailed discussion).