Where does an object trigger an action? An investigation about affordances in space
- First Online:
- Cite this article as:
- Costantini, M., Ambrosini, E., Tieri, G. et al. Exp Brain Res (2010) 207: 95. doi:10.1007/s00221-010-2435-8
- 718 Views
A series of experiments provide evidence that affordances rely not only on the mutual appropriateness of the features of an object and the abilities of an individual, but also on the fact that those features fall within her own reachable space, thus being really ready-to-her-own-hand. We used a spatial alignment effect paradigm and systematically examined this effect when the visually presented object was located either within or outside the peripersonal space of the participants, both from a metric (Experiment 1) and from a functional point of view (Experiment 2). We found that objectual features evoke actions only when the object is presented within the portion of the peripersonal space that is effectively reachable by the participants. Experiments 3 and 4 ruled out that our results could be merely accounted for by differences in the visual salience of the presented objects. Our data suggest that the power of an object to automatically trigger an action is strictly linked to the effective possibility that an individual has to interact with it.
KeywordsAffordancePeripersonal spaceAction representation
The notion of affordance, introduced by Gibson (1979), means the power of the environment to furnish the viewer with possible actions. An affordance is not about a mere physical property, rather it incarnates the action opportunities that the environment may offer to any individual which is able to perceive and use them. Thus, to give some examples, a surface may be stand-on-able or sit-on-able. Similarly, a given object may afford a whole range of motor acts: it may be grasped, thrown, pressed, bitten, kicked, etc.
Many authors have emphasized the mutuality of organism and environment as well as the role of action in defining affordances (Turvey et al. 1981; Shaw et al. 1982; Turvey 1992; Michaels et al. 2001). Notably, it has been argued that affordances should be construed in terms of relationship between the environmental and/or objectual features and the motor abilities of an individual (Chemero 2001, 2003, 2009; Costantini and Sinigaglia in press).
In support of this notion, it has been shown that people’s sensitivity to affordances, particularly to objectual or micro-affordances (Ellis and Tucker 2000), can be measured by means of the compatibility effect. This effect refers to a decrease of reaction times when the subject executes a motor act which is congruent with that afforded by a seen object. For example, observing a handled cup suggests or even demands the movements aimed at reaching for its handle and the grip adequate to grasp it in order to drink (Tucker and Ellis 1998, 2001). In a seminal study, Tucker and Ellis (1998) presented images of everyday graspable objects with handles (e.g. cups). In their study, the objects appeared either right side-up or upside-down, with the handles randomly oriented either to the left or right side of space. The question of interest was whether the movements afforded by the picture of the handle potentiated any form of action. Participants provided left- or right-hand responses to indicate whether the target object was right side-up or upside-down. Tucker and Ellis found a significant compatibility effect—also called spatial alignment effect (Bub and Masson 2010)—based on whether the affording handle appeared on the same side as the responding hand, e.g. left responses were faster when the handle appeared on the left side of space. The evidence reported above has clearly demonstrated that the perception of affordances potentiate specific motor acts that are best suited for manipulating and interacting with the target object.
Given the relational characterization of affordances, the question arises to as whether they are always instantiated by the existence of the objectual features (e.g. the oriented handle of a mug) and the presence of an individual who is able to perceive and to use them, or whether the affordance relations are modulated by the possibility for the objectual features to be actually reached by the individual. Indeed, in everyday life, we act upon objects within a well-defined portion of space—a portion near enough to be reached by our hand. Converging evidence from neurophysiology, neuropsychology and experimental psychology (Rizzolatti and Berti 1990; Gross and Graziano 1995; Rizzolatti et al. 1997; Berti and Frassinetti 2000; Maravita et al. 2003) suggests that there are multiple representations of space, each with its own properties, but in a simplified manner, we can distinguish between three spatial representations originating from the body (Rizzolatti et al. 1997): the space covering the surface of our body (personal space), the space immediately surrounding our body (peripersonal space) and the space that falls far away from our body and it is unreachable by a simple arm movement (extrapersonal space).
Interestingly, the judgment of distance with respect to a given object has been shown to vary according to the action capabilities of the individual (Witt et al. 2005; Witt and Proffitt 2008). In particular, Witt et al. (2005) asked participants to estimate distances to targets as they did or did not hold a tool, with or without the purpose of reaching them. The targets were presented at the same distances in all the conditions. However, the perceived distances were lower when participants not only held the tool but also had the intention to use it in order to reach the presented targets. More recently, Linkenauger et al. (2009) investigated whether and to what extent the perceived distances of an object may be scaled by the difficulty required to pick up an object. Participants were presented with tools with handle orientations that could either facilitate or impede to grasp with their dominant and non-dominant hands. The results showed that right-handed (but not left-handed) participants estimated tools that were more difficult to grasp to be farther than tools that were easier to pick up.
If the perception of the distance with respect to a graspable target is modulated by the easiness to grasp it, what does it happen to the affordance relation when one manipulates the possibility for an individual to reach the object-related features that may evoke a given action such as a grasping action?
To answer this question, we investigated whether objects located either within or outside the perceiver’s reachable space were equally able to potentiate specific motor acts, as measured by the spatial alignment effect.
We used three-dimensional stimuli, which allowed us to give the participants the illusion of objects being located in the near peripersonal space or in the far extrapersonal space. Participants were instructed to replicate a seen grip by performing a reach-to-grasp motor act, with either the right or the left hand, on the presentation of a task-irrelevant go-signal. The go-signal was a 3D scene in which a mug, placed on a table, had the handle oriented toward the left or right, thus being congruent or incongruent with the executed movement. In a first experiment, the mug was located either within the subject’s peripersonal or extrapersonal space. We found spatial alignment effects only when the mug was located within the peripersonal space.
Peripersonal space is usually defined as the space that encompasses the objects within reach (Rizzolatti et al. 1997). According to this definition, it can be construed in two different ways, by putting the emphasis either on the nearness of the object, that is, on its mere distance from the agent, or on the reachability of the object, that is, on its being really reachable.
In a third and a fourth control experiments, we ruled out the possibility that our results were simply driven by the fact that visual saliency of the presented object was reduced when the object was located far away from the participants (as in Experiment 1) or hidden beyond a semi-opaque screen (as in Experiment 2).
Fifteen healthy participants took part in the study (10 women, mean age 25 years, range 22–28). All subjects were right-handed, had normal or corrected-to-normal visual acuity, were naïve as to the purposes of the experiment and gave their informed consent. The study was approved by the Ethics committee of the “G. d’Annunzio” University, Chieti, and was conducted in accordance with the ethical standards of the 1964 Declaration of Helsinki.
Two sets of stimuli were used. The first set of stimuli included colored pictures depicting either a right or a left hand pantomiming of a precision grip movement. The second set of stimuli included 3D scenes. The scenes were 3D rooms, with a table and a mug on it, created by means of 3D studiomax v.13. The handle of the mug was placed either on the right or on the left. In half of the trials, the mug was placed within the participants’ peripersonal space (30 cm; Fig. 1a, left panel) while in the other half in the extrapersonal space (150 cm; Fig. 1a, right panel).
Stimuli were presented by means of a head mounted display with a resolution of 1,024 horizontal pixels by 768 vertical pixels (ProView™, http://www.rockwellcollins.com/optronics). Each trial consisted on the presentation of the instruction stimulus for 150 ms followed, after a variable delay (ranging from 150 to 450 ms), by the go-stimulus lasting 500 ms. Participants were requested to replicate the grip presented in the first set of stimuli (instruction stimuli) by performing a reach-to-grasp motor act as soon as the go-stimulus appeared (Fig. 1e). Thus, congruent trials refer to the condition in which a participant had to pantomime a reach-to-grasp motor act with either the right or the left hand and the handle was located ipsilaterally, while incongruent trials refer to the condition in which the responding hand and the handle were in opposite hemispaces. At the beginning of each trial, participants rested their index fingers on two response buttons arranged horizontally on a button box. Responses were given by lifting the index finger of the response hand and then making the grasping movement as instructed. This allowed us to measure liftoff time (i.e. the time between onset of the go-stimulus and initial hand movement). Each participant provided us with 16 trials per condition. The presentation of the stimuli and the recording of the participants’ responses (in terms of movement onset) were controlled by a custom software (developed by Gaspare Galati at the Department of Psychology, Sapienza Università di Roma, Italy; Galati et al. 2008), implemented in MATLAB (The MathWorks Inc., Natick, MA, USA) using Cogent 2000 (developed at FIL and ICN, UCL, London, UK) and Cogent Graphics (developed by John Romaya at the LON, Wellcome Department of Imaging Neuroscience, UCL, London, UK). At the end of the experiment, participants were requested to judge the distance of the objects with respect to their bodies. Near and far stimuli were judged as being 45 cm (SD = 20 cm) and 150 cm (SD = 5 cm) away, respectively.
Trials in which subjects failed to respond (1.2%) were discarded from the analysis. Error rates did not differ across conditions. The mean RT of the correct responses was calculated for each condition; responses longer than 2 standard deviations from the individual mean were treated as outliers and not considered (3% of the data set). Data were entered in a two-way ANOVA with the Location of the object (peripersonal space vs. extrapersonal space) and Congruency (Congruent vs. Incongruent) as within-subjects factors.
RT analysis revealed a significant Location by Congruency interaction (F1,14 = 8.4, P < 0.05). The interaction was explained by the fact that while RTs to congruent and incongruent trials were comparable in the extrapersonal space, they were significantly higher for incongruent (453 ms) than congruent trials (427 ms, P < 0.05) in the peripersonal space.
Experiment 1 demonstrated that the perception of objectual affordances potentiate motor acts only when the objects are located in the peripersonal space. However, peripersonal space can be defined both in metric and functional terms, and results from Experiment 1 do not allow us to disentangle these two interpretations. To test the hypothesis that what is relevant is not the metric distance between the perceiver and the object but the actual possibility to interact with the presented object, we carried out a second experiment in which we divided the surrounding space of the participants in both a reachable and non-reachable sub-space by presenting the task-irrelevant handled mug in front or beyond a near transparent panel, respectively.
Fifteen healthy participants took part in the study (7 women, mean age 25 years, range 23–28). All participants were right-handed, had normal or corrected-to-normal visual acuity, were naïve as to the purposes of the experiment and gave informed consent. The study was approved by the Ethics Committee of the “G. d’Annunzio” University, Chieti, and was conducted in accordance with the ethical standards of the 1964 Declaration of Helsinki.
Materials and procedure
In this experiment, the first set of stimuli was the same as in Experiment 1, while the second set was different. The mug was always placed within the subjects’ peripersonal space (30 cm). In half of the trials, it was located in front of a transparent plexiglas panel (Reachable Peripersonal Space; Fig. 1b, left panel), whereas in the other half it was located beyond the same panel (Non-Reachable Peripersonal Space; Fig. 1b, right panel). The procedure was the same as in Experiment 1.
Trials in which participants failed to respond (0.6%) were excluded from the analysis. The mean RT of the correct responses was calculated for each condition; responses longer than 2 standard deviations from the individual mean were treated as outliers and not considered (2%). Data were entered in a two-way ANOVA with the Location of the object (Reachable Peripersonal Space vs. Non-Reachable Peripersonal Space) and Congruency (Congruent vs. Incongruent) as within-subjects factors.
Experiment 1 demonstrated that only objects located in the peripersonal space might evoke specific motor acts. In this experiment, however, a possible confound is present. That is, although both the near and the far mugs had the same physical size, the visual angle they subtended change drastically when displayed in these two portions of space. Thus, it can be hypothesized that the larger visual angle subtended by the near compared to the far mug had an impact on the evocation of the grasp response. To control for this possible confound, we ran a third experiment in which the mug had the same angular size at different nominal distances. Thus, the far cup was phenomenally larger than the near cup.
Fifteen healthy participants took part in the study (11 women, mean age 24 years, range 22–27). All subjects were right-handed, had normal or corrected-to-normal visual acuity, were naïve as to the purposes of the experiment and gave informed consent. The study was approved by the Ethics Committee of the “G. d’Annunzio” University, Chieti, and was conducted in accordance with the ethical standards of the 1964 Declaration of Helsinki.
Materials and procedure
The first set of stimuli was the same as the previous experiment. The second set of stimuli included 3D scenes. The scenes were 3D rooms, with a table and a mug on it, created by means of 3D studiomax v.13. The handle of the mug was placed either on the right or on the left. In half of the trials, the mug was placed within the participants’ peripersonal space (30 cm; Fig. 1c, left panel), while in the other half in the extrapersonal space (150 cm; Fig. 1c, right panel). It is worth noting that both the near and the far mugs subtended the same visual angle (2 × 1.4 degrees). This was done in order to avoid differences in visual inputs between the two sectors of space. Notwithstanding the differences in size, both the mugs were judged as compatible with the movement to be executed. The procedure was the same as in the previous experiments.
Trials in which participants failed to respond (1.2%) were excluded from the analysis. The mean RT of the correct responses was calculated for each condition; responses longer than 2 standard deviations from the individual mean were treated as outliers and not considered (3%). Data were entered in a two-way ANOVA with Location of the object (Peripersonal vs. Extrapersonal space) and Congruency (Congruent vs. Incongruent) as within-subjects factors.
RT analysis revealed a significant Location by Congruency interaction (F1,14 = 5.1, P < 0.05). The interaction was explained by the fact that while RTs to congruent and incongruent trials were comparable in the Extrapersonal space, they were significantly higher for incongruent (415.7 ms) than congruent trials (382.1 ms, P < 0.05) in the Peripersonal space (see Fig. 2).
Experiment 2 clearly indicated that the perception of objectual affordances potentiate specific motor acts, as measured by the spatial alignment effect, only when the target-objects are located within the actually reachable space. However, an alternative explanation might also account for our results. That is, when the object is hidden behind a semi-opaque screen, participants need more time to extract the affordance-inducing feature than they need to initiate their pre-planned response. To control for this possible confound, we conducted a further experiment in which the mug within the visual scene was either clearly visible or masked with a spatial filter. In this experiment, we also aimed at better characterizing the compatibility effect found in the previous experiments, in terms of inhibition on incongruent trials and/or facilitation on congruent trials. To this end, we included a neutral no-handle trials in order to compare congruent and incongruent conditions.
Fifteen healthy participants took part in the study (8 women, mean age 24 years, range 22–28). All participants were right-handed, had normal or corrected-to-normal visual acuity, were naive as to the purposes of the experiment and gave informed consent. The study was approved by the Ethics Committee of the “G. d’Annunzio” University, Chieti, and was conducted in accordance with the ethical standards of the 1964 Declaration of Helsinki.
Materials and procedure
In this experiment, the first set of stimuli was the same as in the previous experiments, while the second set was again different. The mug was always placed within the participant’s peripersonal space (30 cm). The handle could be oriented toward left, right or hidden beyond the mug (no-handle neutral condition). The mug was clearly visible in half of the trials (see Fig. 1d, left panel), whereas in the other half a spatial filter was applied on it. The latter stimuli were produced in Adobe Photoshop (Version CS3) by applying a “mosaic” pixelation filter, using a value of 7 pixels to resample the images at a lower resolution (see Fig. 1d, right panel). Although at lower resolution, the mug was clearly recognizable by all the participants as self reported. The procedure was the same as before.
Trials in which participants failed to respond (1%) were excluded from the analysis. Outliers (3%) were computed as earlier. Data were entered in a two-way ANOVA with the Visual Salience of the object (High resolution vs. Low resolution) and Congruency (Congruent vs. Incongruent vs. Neutral) as within-subjects factors.
RT analysis revealed a significant main effect of Congruency (F2,28 = 4.43, P < 0.05) showing how RTs to congruent trials (383.7) were faster than RTs to incongruent trials (418.0; P < 0.05). Moreover, RTs to congruent trials (383.7) tended to be faster than RTs to neutral trials (406.8; P = 0.059, Fig. 2). Both the main effect of Visual Salience and its interaction with Congruency were not significant.
According to Gibson (1979), affordances are properties of the environment providing the viewer with practical opportunities which she is able to perceive and use. Post-Gibson attempts to highlight what affordances are have mainly emphasized the mutuality of organism and environment in defining them (Turvey et al. 1981; Shaw et al. 1982; Turvey 1992; Michaels et al. 2001). Notably, it has been proposed to construe affordances in terms of a specific relationship between the environmental and/or objectual features and the motor abilities of an individual (Chemero 2001, 2003, 2009; Costantini and Sinigaglia in press).
Here, we refine the notion of affordance by providing empirical data showing that the affordance relation, at least in the case of a situated object, is based not only on the mutual appropriateness of its constituent relata (i.e. objectual features and individual’s motor abilities) but also on their spatial relationship. Indeed, in the first Experiment, we found that an objectual feature (i.e. the oriented handle of a mug) might suggest or even demand a specific motor act (i.e. a reach-to-grasp motor act) only when the object is presented within the observer’s own peripersonal space. In the second Experiment, we further expanded these findings by demonstrating that the functional characterization of the peripersonal space is more relevant for the affordance relation than its mere metric description. Indeed, we found that the objectual feature might suggest or even demand a specific motor act only when the object is actually reachable by the observer, that is, when it is literally ready-to-her-own-hand.
Possibly, one may argue that the way in which visual stimuli were presented did not provide participants with a realistic three-dimensional immersive experience, thus lessening the relevance of our results. However, although our experimental setup did not take advantage of all the crucial elements of virtual reality, it allows us to selectively manipulate the actual reachability of the presented affording features and to reliably assess the relationship between reaching space and affordance.
Our data appear to be fully consistent with the notion that the peripersonal space evolved primarily to subserve action (Rizzolatti et al. 1997). Several neurophysiological (Gentilucci et al. 1988; Fogassi et al. 1992; Graziano et al. 1994) and neuropsychological studies (Rizzolatti and Berti 1990; Gross and Graziano 1995; Rizzolatti et al. 1997; Berti and Frassinetti 2000; Maravita et al. 2003) demonstrated that the peripersonal space is motor in nature, its extent depending on the varying range of the individual’s reaching abilities. Indeed, single cell recordings from the inferior parietal lobule of the macaque brain showed that the extent of the peripersonal space can be modified by tool use (Iriki et al. 1996). Similar results have been also found in healthy (Maravita et al. 2002; Serino et al. 2007) and brain damaged humans (Farnè et al. 2005, 2007; Neppi-Mòdona et al. 2007). In particular, line-bisection studies on patients with selective neglect for peripersonal space showed that tool use might reduce or increase the severity of the neglect (Berti and Frassinetti 2000; Pegna et al. 2001; Ackroyd et al. 2002; Neppi-Mòdona et al. 2007). Likewise, tool use in patients with visuo-tactile extinction has been demonstrated to modulate the severity of their extinction (Farnè and Ladavas 2000; Farnè et al. 2007).
Interestingly, the only study that explicitly investigated the relationship between affordance and space failed to find any interaction (Tucker and Ellis 2001). A possible interpretation for these apparent conflicting results can be found in the different task performed by the participants in the two studies. Indeed, Tucker and Ellis (2001) asked participants to signal whether an object was natural or manufactured by grasping with a precision grip or with a whole hand prehension a response device, respectively. In different blocks, objects could be located at 15 or 200 cm far away from the participants. It is worth nothing that such a categorization task was accomplished without any effective reaching movement, as the response device was already held by the participants and they had just to change the grip.
In contrast, in our study, although there was not a real target object, participants had to actually execute a reach-to-grasp motor act. There is a large evidence that, although the reaching and grasping movements are strictly intertwined, their planning and execution depend on different and segregated neural processes (Jeannerod 1988; Jeannerod et al. 1995). Indeed, grasping an object with the hand requires to transform the visual features of the object into the appropriate shaping of the fingers as well as to control the execution of their movements. Space localization is critical for reaching only: its successful planning and execution depends on the processing of the spatial relationship between the object to be reached and the body parts involved in the reaching movements (Rizzolatti and Sinigaglia 2008).
Thus, it is likely that our effective reach-to-grasp motor task was able to reveal the interaction between affordance and space because, differently from the Tucker and Ellis’s (2001) task, not only was sensitive to the information on the objectual features but also automatically recruited the representation of the participants’ own reachable space.
One might argue that objectual features either located far away from the observer (Experiment 1) or hidden beyond a semi-opaque screen (Experiment 2) are extracted later than the time needed to initiate a pre-planned response. This could have potentially biased our results by eliminating the spatial alignment effect in these two conditions. However, Experiments 3 and 4 clearly indicate that this is not the case. In Experiment 3, we showed that the size alone had no impact on the evocation of the grasp response, while in Experiment 4 we ruled out the possibility that degraded stimuli were not able to trigger proper actions. Indeed, when augmenting the visual salience of the far object (Experiment 3) or reducing the visual salience of the near object (Experiment 4), the spatial alignment effect was not abolished.
A final issue that needs to be addressed is the nature of the motor processing affected by our manipulation. Participants were instructed to replicate a reach-to-grasp motor act after a delay, thus they prepared it and executed it as soon as a go-signal appeared (see also (Stins and Michaels 1997). Incidentally, the go-signal activated a motor schema that was either congruent or incongruent with the planned motor act. Imitation research has revealed motor facilitation effects when participants respond according to some property of an observed action. Responses are usually faster when the response and the observed actions are congruent rather than incongruent. Similar motor facilitation has been observed regardless of whether an individual views a static hand (Craighero et al. 2002; Vogt et al. 2003; van Schie et al. 2008) or a dynamic presentation of a grasping hand (Brass et al. 2000; Sturmer et al. 2000). Also our spatial alignment effect was due to a facilitation, as demonstrated by congruent trials being faster than neutral trials in Experiment 4. We suggest that the reaching- and grasping-related motor programs (triggered by the movement to be imitated) and the object-related motor programs (triggered by the visual presentation of the object within the participants’ reaching space), when simultaneously and congruently activated do facilitate each other.
To conclude, our findings suggest that, at least for the objectual or micro-affordances (Ellis and Tucker 2000), the relation of affordance turns out to be space constrained, as it depends on the spatial relationship between its relata, namely the objectual features and the individual’s motor abilities. Although such a spatial relationship is not constitutive of the distinctiveness of the affordances (e.g. the graspability of an object as such), it makes them possible. Indeed, affordance relations are not always instantiated whenever their relata appear to be mutually appropriate, but only when the objectual features occur in the reachable space of an individual endowed with the appropriate motor abilities, i.e. when they are really ready-to-her-own-hand.
CS was supported by a grant from San Paolo Foundation and by a grant from MIUR (PRIN 2007). GC was supported by grant from MIUR (PRIN 2007). The authors wish to thank the referees and Vittorio Gallese for their insightful comments on a previous version of the manuscript.