The allocation of valenced concepts onto 3D space

ABSTRACT

The valence–space metaphor research area investigates the metaphorical mapping of valenced concepts onto space. Research findings from this area indicate that positive, neutral, and negative concepts are associated with upward, midward, and downward locations, respectively, in the vertical plane. The same research area has also indicated that such concepts seem to have no preferential location on the horizontal plane. The approach–avoidance effect consists in decreasing the distance between positive stimuli and the body (i.e. approach) and increasing the distance between negative stimuli and the body (i.e. avoid). Thus, the valence–space metaphor accounts for the mapping of valenced concepts onto the vertical and horizontal planes, and the approach–avoidance effect accounts for the mapping of valenced concepts onto the “depth” plane. By using a cube conceived for the study of allocation of valenced concepts onto 3D space, we show in three studies that positive concepts are placed in upward locations and near the participants’ body, negative concepts are placed in downward locations and far from the participants’ body, and neutral concepts are placed in between these concepts in both planes.

KEYWORDS:

• Metaphorical mapping
• embodied cognition
• valence–space metaphor
• approach–avoidance effect
• the cognition cube

Acknowledgements

FM-R thanks Daniela Álvarez Lleras for the illustrations of the experimental device and the experimental sequence, Susan Brunner and Rosie Gronthos for proofreading the manuscript, and Alexandra Marmolejo-Losyeva (mañuña) and Iryna Losyeva for their permanent support.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Fernando Marmolejo-Ramos http://orcid.org/0000-0003-4680-1287

Notes

1. Most studies in approach–avoidance behaviours have relied on joystick tasks that are confounded by the semantic labels of movement responses and the spatial relation between the participant and the joystick (see Seibt, Neumann, Nussinson, & Strack, 2008). Krieglmeyer et al. (2010), however, provide evidence that favours an automatic association between positive–negative stimuli and distance-changing consequences. Note that participants in the current experiment could have formed intentions/decisions prior to pointing to the location of where the word would be placed (see details in the “Methods” section below). Thus, decisional processes, and associated processes of intention formation (action coding), could have played a significant part in the results.

2. The sample size for each experiment was determined prior to data collection for a linear model whose five predictors explained at least 25% of the variability in the dependent variable (i.e. f² = 0.33; a large effect size) with an alpha of 0.05 and a power of 80%. The estimated sample size was 44 (44.35 in fact). Although we did not reach that number, a sample size of 40 would provide an observed power (OP) of 74.21% (f² = 0.33); a good-enough power not far from 80%. These estimations were achieved via the R function “pwr.f2.test” implemented in the “pwr” package; where u is the degree of freedom in the numerator (i.e. the number of predictors), v is the degree of freedom in the denominator (note v = n − u − 1; where n is the sample size), and f² is the effect size (f² = R²/(1 − R²); where R² is the variance accounted for). Hoenig and Heisey (2001) show the one-to-one relationship between OP and p-values such that significant p-values always correspond to a high OP (in other words, “OP returns a p-value under a different disguise” D. Cousineau, personal communication, 13 December 2016). Thus, no OPs are reported herein. For meta-analytical purposes, the p-values from independent studies can be combined (e.g. via the Stouffer method; see also the “metap” R package).

3. The robust models used in the analyses were all sequential and fixed-effects. The results echoed those reported here when the data were submitted to a fully fixed-effects simultaneous model. Data were also submitted to a linear mixed-effects models in which participants and items were entered as random effects (note in Experiment 1 there were only three items; thus these were treated as fixed effects). The participants factor was not significant in Experiment 1 in any of the three dependent variables; participants and items were significant in Experiment 2 for data in the Y axis only; and only the participants factor was significant in the Y and Z axis data in Experiment 3.

4. It would have been more informative to assess the effect of handedness and the interaction of this factor with that of gender. However, the large disparity between the number of right- and left-handed participants and the reduced number of left-handers precluded such analyses (see Table 1). After excluding the left-handed participants in each experiment (Experiment 1: 2.5%, Experiment 2: 5%, and Experiment 3: 10%), the pattern of results reported herein remained.

5. It is important to note that this association was only found in a task block in which the movement of the manikin was congruent with instructed movement goals (see Eder, Rothermund, & Hommel, 2016, for a detailed discussion). Furthermore, approach–avoidance studies involving manual actions have typically failed to find approach-avoidance effects in non-evaluative categorisation tasks (for a recent meta-analysis see Phaf et al., 2014).

6. Phonetically inward words are associated with consonant wonderings from the front to the rear of the mount (i.e. swallowing- or deglutition-like movements), while outward words are associated with consonant wonderings from the rear to the front of the mouth (i.e. spitting- or expectoration-like movements). The underlying idea is that the phonetics of inward words leads to an orofacial movement resembling an approach movement, whereas the phonetics of outward words resembles an avoidance movement.

7. It would be interesting to employ inward and outward words (as those used by Topolinski et al., 2014) in lieu of the valenced adjectives we used. It could be hypothesised that people would tend to place inward words more closely to themselves than outward words.

Additional information

Funding

The study was supported by the Swedish Foundation for Humanities and Social Sciences [grant number M14-0375:1]. This work was also supported by funds to AA from the International Postdoc grant from the Swedish Research Council [grant number VR2014-240].