Good is up—spatial metaphors in action observation

Gottwald, Janna M.; Elsner, Birgit; Pollatos, Olga

doi:10.3389/fpsyg.2015.01605

ORIGINAL RESEARCH article

Front. Psychol., 20 October 2015

Sec. Emotion Science

Volume 6 - 2015 | https://doi.org/10.3389/fpsyg.2015.01605

Good is up—spatial metaphors in action observation

$\r\nJanna M. Gottwald*$ Janna M. Gottwald¹^*

Birgit Elsner²

Olga Pollatos³

¹Department of Psychology, Uppsala University, Uppsala, Sweden
²Department of Psychology, University of Potsdam, Potsdam, Germany
³Department of Health Psychology, Ulm University, Ulm, Germany

Positive objects or actions are associated with physical highness, whereas negative objects or actions are related to physical lowness. Previous research suggests that metaphorical connection (“good is up” or “bad is down”) between spatial experience and evaluation of objects is grounded in actual experience with the body. Prior studies investigated effects of spatial metaphors with respect to verticality of either static objects or self-performed actions. By presenting videos of object placements, the current three experiments combined vertically-located stimuli with observation of vertically-directed actions. As expected, participants’ ratings of emotionally-neutral objects were systematically influenced by the observed vertical positioning, that is, ratings were more positive for objects that were observed being placed up as compared to down. Moreover, effects were slightly more pronounced for “bad is down,” because only the observed downward, but not the upward, action led to different ratings as compared to a medium-positioned action. Last, some ratings were even affected by observing only the upward/downward action, without seeing the final vertical placement of the object. Thus, both, a combination of observing a vertically-directed action and seeing a vertically-located object, and observing a vertically-directed action alone, affected participants’ evaluation of emotional valence of the involved object. The present findings expand the relevance of spatial metaphors to action observation, thereby giving new impetus to embodied-cognition research.

Introduction

The ups and downs of life. In everyday speech many metaphors link vertical space and emotional valence. People, objects, and actions with positive attributes are associated with metaphors pertaining to their physical highness, while those perceived negatively are often related to physical lowness. These fundamental concepts are thought to develop early in life—to some extent in the preverbal period (Lakoff and Johnson, 1980; Barsalou, 2009). Accordingly, metaphors are more than a linguistic phenomenon. From an embodied-cognition point of view metaphors are grounded in our bodies and therefore in our physical experience. In particular, spatial metaphors like “good is up” and “bad is down” are inseparable from their experiential basis and they enable us to orient within our world—they are the “heart of our conceptual system” (Lakoff and Johnson, 1999, p. 30; Glenberg and Kaschak, 2002; Casasanto, 2009; Shapiro, 2011).

Despite of the often-mentioned assumed embodied basis of spatial metaphors, much of the previous research on the relations between metaphors of vertical space and emotional valence has focused on the evaluation of static stimuli located within different vertical positions. One approach asked whether vertical position affects ratings of emotional valence for neutral stimuli. In their study on the connection between verticality and perceived physical attractiveness, Meier and Dionne (2009) presented several images of females and males, who scored similarly on attractiveness ratings, in various vertical positions. The expected results occurred only when participants saw images of their own sex: Participants rated the up-presented persons as more attractive than the down-presented persons. Similar results were found when participants rated up- versus down-presented images of persons in terms of higher versus lower religiousness as a positive value (Meier et al., 2007).

Another study examined whether vertical position influences (spatial) memory or attention for emotional stimuli. Crawford et al. (2006) found a typical bias, in that their participants remembered positive images from the international affective picture system (IAPS) as being located at higher screen positions, and negative images at lower positions, than they had been de facto. Likewise, in a study by Meier et al. (2007) on the association between vertical space and the divine, participants remembered “god”-like images metaphor-consistently as appearing higher, and “devil”-related images as appearing lower than neutral control pictures. In a categorization task, participants encoded “god”-related words (such as “Almighty”) faster if these were in a high vertical position as opposed to the low vertical position and “devil”-related words (such as “Lucifer”) faster if these were in a low vertical position as opposed to the high vertical position. These findings corroborate with a previous study in which Meier and Robinson (2004) let their participants evaluate affective words as either positive or negative. They found that the ratings were faster if the words were presented metaphor-consistently in vertical space.

While considerable body of work has demonstrated relations between vertical position and emotional valence, few studies have investigated the effect of metaphors by allowing individuals to act in the vertical dimension of space. For instance, Casasanto and Dijkstra (2010) studied the interaction between emotional memory and motion: While moving their hands up or down (by moving marbles), participants should tell a pleasant or an unpleasant past episode of their life. Participants were faster at retrieving memories if their movements and memory valence were congruent (e.g., moving marbles up and telling a positive episode), and they were also more likely to retrieve a memory whose valence was congruent with their movement.

Seno et al. (2013) built on these results and made use of illusory self-motion to further distinguish between effects of motion perception concerning the observer and concerning the observed stimulus. Actually stationary observers were presented with upward- or downward-moving patterns (inducing the reversed self-motion experience), and were asked to retrieve autobiographical episodic memories. Participants more likely recollected positive or negative memories in congruency with their illusory upward or downward self-motion direction, i.e., incongruent with the motion direction of the pattern. Thus, it is one’s perceived motion direction and not the actual motion direction of the stimulus that drives the mood-congruency effect, which highlights the importance of the body in metaphor-related effects.

All previous studies have investigated the relationship between metaphors of vertical space and emotional valence either by having participants look at objects located in vertical space, or allowing them to place the objects themselves. Combining these aspects is an interesting approach, especially because we learn various motoric behaviors and gain many of our experiences while observing others interacting with objects (Mattar and Gribble, 2005; Elsner, 2009). The present research sought to investigate whether observing a person acting on an object in vertical space can also trigger the metaphor “good is up” (respectively “bad is down”) by presenting the objects in the context of dynamic actions mimicking real-life experiences. Specifically, the current work asked whether evaluations of emotional valence of everyday objects are more positive when the objects are observed being placed up (i.e., higher in vertical space) as compared to when they are observed being placed down (i.e., lower in vertical space).

Study 1

In an initial study, we used two vertical positions (“up” and “down”). We based our hypotheses on effects of the vertical positions and expected that emotionally neutral objects would be rated more positively in terms of emotional valence when they are observed being placed up in vertical space as compared to when they are observed being placed down. It is important to note that we expected effects of vertical positioning only for ratings that involve emotional valence (e.g., gift likelihood, willing to pay), but not for ratings related to the cover story, i.e., ratings aimed at object physicality (i.e., attractiveness of color or shape).

Method

Participants

The sample consisted of 41 German university students (age M = 22.71, SD = 2.75; all female and right-handers) who received a course credit for participation. Participants were recruited with announcement advertising a study on product evaluation. None of the participants was aware of the purpose of the experiment. Approval from the local ethical boards was obtained in accordance to the Declaration of Helsinki.

Stimuli

Videos showing the placement of an emotionally neutral object in vertical space served as stimuli: A woman placed a cup in a wooden bookcase with three vertically positioned shelves (Figure 1A). The woman almost filled the left half of the screen, and the bookcase filled the right half. The positions of the shelves were defined in relation to the woman’s body. The lowest shelf (“down”) was on the same height as her elbow, when her arm was to her side. The “middle” shelf was the same height as her chest, while the highest shelf (“up”) was the same height as her chin. In order to create similar conditions with regard to motor effort, the distances between the shelves were of equal size, and they were determined in relation to the length of the woman’s forearm.

FIGURE 1

Figure 1. Screen shots of the videos used in Study 1 (A), Study 2 (B), and Study 3 (C). Starting position (left), end of the upward/downward movement (middle), and end positions of the placements (right; Study 1: up, down; Study 2: up, middle, down; Study 3: full-screen picture of the cup, spatially neutral position).

Twenty-four cups, differing in form, color, design, and two placements (“up” and “down”) were used to create 48 video stimuli. Every cup was shown in each position to rule out the influence of possible preferences for certain cups. Each video took 6 s and started with showing the woman in profile holding a cup in her right hand, in height of the middle position. The woman looked first at the bookcase and then at the cup in her hand. She then moved her right arm and placed the cup on one of the shelves (“up” or “down”) and held the cup for the last 2 s of the scene. Her movement was such that she did not obstruct the cup’s visibility to the observer with her hand or her arm.

Procedure

Participants were seated in an experimental cabin and watched the videos. They first saw two training trials, using videos with a cup that was not in the final sample. Then, the 48 stimulus videos were shown on an 18-inch screen to every participant in a randomized order. Prior to the presentation of each video, a 50 Hz tone and a centered fixation cross were presented. After each video, six items demanding ratings of different aspects of the presented cup, including two main target variables of emotional valence, were presented each after another on the screen (i.e., 288 ratings in total).

The first three items were aimed at object physicality underlining the cover story that this study was about product evaluation. Item 1 was about shape (“How attractive do you find the shape of the cup?”) and item 2 was about color (“How attractive do you find the color of the cup?,” English equivalents). These two items were later combined into one mean rating of Esthetics. Item 3 (“How do you appraise the usability of the cup?”) was not evaluated. The next two items were target items of emotional valence being associated with the vertical placement of the object: “How likely is it that you would choose this cup as a gift for your best female friend?” (item 4: Gift likelihood) and “How much would you pay for this cup?” (item 5: Willing to pay; cf. Lee and Schwarz, 2010). Item 6 asked explicitly about emotional valence: “How much do you like this cup?” (Liking).

Below each question, a horizontal visual analog scale was presented. Scales ends ranged from 1 (items 1, 2, 3, 4, 6: not at all; item 5: 0.99€) to 10 (items 1, 2, 3, 4, 6: very much; item 5: 4.59€). For item 5, amounts of money were specified, increasing by 0.40 € across each of the 10 steps. Participants answered by pressing a key on a German keyboard and were instructed to answer as quickly and as accurately as possible. The middle line of letter keys was used, from “A” = “1” to “Ö” = “10”. After the experiment, participants were asked about the assumed purpose of the experiment. Presented with two options “product evaluation” and “other” (that had to be defined by a free answer), they could either tick one or both options. The experiment took about 30 min in total.

Data Analysis

The average ratings for every cup in each spatial position were calculated and averaged across all cups. Then, separate analyses of variance (ANOVAs) for the four aspects (i.e., Esthetics, Gift likelihood, Willing to pay, Liking) were conducted with the within-subjects factor spatial position (up, down).

Results

Testing of normal distribution of the data was performed using Shapiro–Wilk tests and revealed no violation of normality. There was a significant main effect of spatial position for Gift likelihood, F(1,40) = 6.48, p = 0.015. As hypothesized, Gift likelihood was higher for objects observed to be placed up than for objects observed to be placed down, MD (mean difference) = 0.48, SD = 0.19, r = 0.14. Hence, cups that were placed on the higher shelf were more likely to be chosen as possible gifts than were cups placed on the lower shelf (see Figure 2A).

FIGURE 2

Figure 2. Mean ratings of Gift likelihood (choosing the cup as a gift for a friend), Willing to pay (amount of money), and Esthetics (attractiveness of the cup’s shape or color) as a function of the cups’ vertical placement “up” versus “down” in Study 1 (A), Study 2 (B), and Study 3 (C), as well as “middle” in Study 2 (B). Error bars indicate standard errors. Scale: 1 = not at all/0.99€, 10 = very much/4.59€.

The main effect of spatial position for Willing to pay was also significant, F(1,40) = 5.52, p = 0.024. As expected, the indicated amount of money was higher for objects observed to be placed up as compared to down, MD = 0.56, SD = 0.34, r = 0.12. Thus, participants would have paid more for cups that were placed on the upper shelf than for cups placed on the lower shelf.

There also was a significant main effect of spatial position on the item asking explicitly about Liking, F(1,40) = 8.21, p = 0.007. Participants liked the cups that were observed to be placed up more (M = 4.78; SD = 1.68) than those that were observed to be placed down (M = 3.98; SD = 1.19), MD = 0.81, SD = 0.30, r = 0.17.

As expected, there was no significant main effect of spatial position for Esthetics (mean scores across items 1 and 2), F(1,40) = 0.32, p = 0.572. Ratings concerning the cup’s shape or color did not differ significantly for the observed vertical placements, MD = 0.44, SD = 0.57, r < 0.01.

It has to be noted that the reported mean ratings were in a rather low range compared to possible scores up to 10 (see Figure 2A). The differences between ratings across the conditions and the effect sizes were small (cf. Cohen, 1992).

The systematic question about the assumed purpose of the experiment revealed that none of the participants was aware of the purpose. Most participants believed the study was only about “product evaluation” (90.24%); few thought it was about “product evaluation” and “drinking behavior” respectively “thirst” (7.32%); one participant believed the study was only about “thirst” (2.44%).

Discussion

The present study tested the assumption that observation of a vertically-directed action affects the evaluation of emotional valence of an object via influence of the spatial metaphors “good is up” or “bad is down.” Indeed, main results show that observing a placement of an emotionally neutral object leads to a more positive rating of emotional valence when the object is placed up as compared to down.

Our hypotheses were supported by the results. As expected, we did not find any differences between the vertical positions in participants’ ratings for the two cover-story items about object physicality (i.e., the object’s shape or color). However, vertical position had the expected significant effects on all three target items of emotional valence. Although it has to be noted that the effects were small, participants rather chose a cup as a gift for a friend, were willing to pay more, and explicitly liked the cup better when this cup was observed being placed up as compared to being placed down. Hence, measuring emotional valence directly did not minimize the effects, but revealed similar effects compared to more indirect measurements (as Gift likelihood or Willing to pay). In sum, Study 1 extended the results of previous research on the effects of spatial metaphors on ratings of emotional valence (e.g., Meier et al., 2007; Meier and Dionne, 2009) to observed vertically-directed actions and vertically-located end positions of objects.

Study 2

Given the small, but significant effects of Study 1, we aimed at replicating our findings and at strengthening our initial interpretation in Study 2. Moreover, to investigate whether an additional differentiation in vertical space interacts with emotional valence and to examine the specific effects of vertical placement, we used three vertical positions (up, middle, down). This was done because just comparing “up” versus “down” does not inform about whether each positioning actually changes emotional valence compared to a neutral condition (i.e., middle). Because we did not find different effects of vertical position on explicit or implicit ratings of emotional values in Study 1, we used only the two more implicit items of emotional valence in Study 2 (i.e., Gift likeliness, Willing to pay). Again, we expected first, that emotional valence of the cups would be rated more positively when cups are observed being placed up, as compared to down, in vertical space. Second, we expected that cups observed being placed “up” (spatially higher) should be rated more positively than those being placed in the “middle” position. Third, we hypothesized that cups observed being placed “down” (spatially lower) should be rated less positively than those being placed “middle.” As in Study 1, we expected effects of vertical positioning only for ratings that involve emotional valence (i.e., Gift likeliness, Willing to pay), but not for ratings related to the cover story (i.e., Esthetics).