The Earth Is Our Home

The Earth is our Home:

Systemic Metaphors to Redefine our Relationship with Nature

Paul H. Thibodeau

Cynthia McPhersonFrantz

Matias Berretta

1. Identifying Metaphors

A corpus of metaphors for the human-environment relationship was developed primarily from the existing research literature (e.g., Lakoff 2010; Larson 2011; Nerlich, Koteyko, & Brown 2010; Princen 2010). For example, Princen (2010; Tables 1 and 2) identifies nine metaphors that seem to separate humans from nature (that describe the earth as a “machine,” “laboratory,” “bank,” “store,” “park,” “frontier,” “threat,” “commons” and “colony”) and 10 metaphors that seem to connect humans to nature (that describe the earth as a “spaceship,” “planet,” “watershed,”“scale,” “seed,” “network,” “tide,” “homestead,” “gift,” and “national banking system”). We also consulted colleagues and considered how the human-environment relationship is depicted in popular media in developing our initial pool of metaphors. For example, the earth is commonly anthropomorphized in religious texts and in the popular media, as a “mother” or “child” (Primavesi 2008).

Building off prior linguistic analyses (e.g., Lakoff 2010; Larson 2011; Nerlich, Koteyko, & Brown 2010; Princen 2010), we grouped the metaphors for the human-environment relationship into five classes. That is, in the research literature and in our discussions with colleagues, we found that there were roughly five categories of metaphors commonly used to describe the human-environment relationship: that personify the earth (e.g., as our “mother”),that mechanize the earth (e.g., as a “spaceship”), that highlight the earth’s resources (e.g., as a “store”), that situate the earthspatially (e.g., as a “park”), and that describe the earth as an emergent entity (e.g., as a “community”).

Several considerations influenced our selection of the 17 metaphors that were tested in Study 1:

1. We included at least one metaphor from each of the five categories (i.e. metaphors that personified, mechanized, and commodified, the earth, as well as metaphors that situated the earth spatially and as an emergent entity).
2. We identified metaphors that could be interpreted by participants in nominal form (i.e. using a noun to categorize the earth, as in “The earth is an X”). For example, a “commons” metaphor, that highlights the tragic nature of joint ownership, has been used to frame the relationship between humans and the environment (Princen 2010). However, we did not expect participants to be able to grasp this interpretation without additional context, so it was not included in Study 1.
3. We did not include phrases that seemed to lack a clear figurative interpretation. For example, the earth is commonly described as a “planet,” which has been discussed as a metaphor for the human-environment relationship (Princen 2010). However, we did not expect participants to interpret the sentence, “The earth is a planet,” as a metaphor, so it was not included in Study 1.
4. We included metaphors that have been identified in the research literature as likely to promote a disconnected conception of the human-environment relationship (e.g., “The earth is a machine”), as well as metaphors that have been identified in the research literature as likely to promote a more connected conception of the human environment relationship (e.g., “The earth is a network”; Princen 2010).
5. Finally, when we were unsure of whether a metaphor met the criteria listed above or if two metaphors seemed very similar to one another, we erred on the side of inclusion. For example, both “parent” and “mother” metaphors anthropomorphize the earth and situate humans as direct offspring. Instead of using our own intuition to choose which version to include, we tested both of these metaphors.

With these considerations in mind, 17 metaphors for the human-environment relationship were tested in Study 1. As noted in the main text, the stimulus set included four metaphors that personified the earth (mother, parent, child,ancestor),three that mechanized the earth (spaceship, machine, network), six that highlighted the earth’sresources (bank, store, market, farm, investment, and gift), and four that situated the earthspatially(park)and/orasanemergentpropertyofagroup(kingdom,community,andhome). We used the results of Study 1 to select a subset of six metaphors for further testing in Study 2, as discussed below and in the main text of the paper.

2. Study 1: General Public

2.1 Methods

Participants from the general public were recruited from Amazon’s Mechanical Turk. They were asked 10 ratings questions of 17 metaphors and to identify who they thought would be likely to use the metaphor. The order in which participants evaluated the 17 metaphors was randomized across participants. After rating the 17 metaphors, participants were asked to choose a favorite metaphor, a least favorite metaphor, and to explain these two choices. The instructions for this section read:

People often use analogies or metaphors to talk about the way humans interact with the natural world. These help us understand the way we relate to the environment by comparing nature to something more familiar and concrete, even if that something might not be directly related to the natural world.

On the following screens we will ask you to consider some of these metaphors and to make judgments about them. Please do your best and be honest. There is no right or wrong answer to these questions. Your thoughts are very valuable to us and we greatly appreciate your time and energy.

The ratings questions and scales used to evaluate the metaphors are shown in Table S1.

Table S1. Target questions asked about each metaphor in Study 1 with scales below (bolded questions asked of climate experts in Study 2).

Dimension and
Label / Question and Scale
AFFECT
Like / How much do you like this metaphor?
I hate this metaphor (1) to I love this metaphor (7)
Thought / How thought provoking is this metaphor for you?
Not thought provoking at all (1) to Extremely thought-provoking (7)
Familiarity / How familiar does this metaphor seem to you?
Not familiar at all (1) to Extremely familiar (7)
SYSTEMIC
Complexity / Does this metaphor make the earth seem more complex or more simple?
Entirely simple (1) to Extremely complex (7)
Connected / Does this metaphor make you feel more or less connected to the environment?
Extremely disconnected (1) to Extremely connected
Influence / How much does this metaphor make you feel that humans affect the environment?
Not at all (1) to A great amount (7)
Affected / How much does this metaphor make you feel that the environment affects humans?
Not at all (1) to A great amount (7)
Dynamic / Does this metaphor make you feel as though the earth is static and unmoving, or that it is constantly changing?
Entirely static (1) to Entirely changing (7)
ACCURATE
Accurate / How accurate does this metaphor seem to you?
Not accurate at all (1) to Extremely accurate (7)
Misleading / How misleading does this metaphor seem to you?
Not misleading at all (1) to Extremely misleading (7)

Participants were then asked to complete six attitudinal questionnaires (see Table S2): the revised New Ecological Paradigm (rNEP; 15 items; Dunlap, van Liere, Mertig, & Jones 2000), the Connectedness to Nature Scale (CNS; 5 items; Mayer & Frantz 2004), a measure of Free Market Ideation (3 items; Heath & Gifford 2006), the Conspiricist Ideation Scale (4 items; Lewandowsky, Gignac, & Oberauer 2013), the Systems Thinking Scale (15 items; Davis & Stroink 2016), and a measure of Religiosity (5 items; Strawbridge, Sherma, Cohen, Roberts, & Kaplan 1998).

Table S2. Attitudinal questionnaires completed by participants in Study 1.

Measure / Example Item and Scale
rNEPa (α = .86) / Humans are severely abusing the environment.
(15 items) / Strongly disagree (1) to Strongly agree (5)
CNSb (α = .92) / I often feel a strong connection to nature.
(5 items) / Strongly disagree (1) to Strongly agree (5)
FMIc (α = .67) / The free-market system works equally well in both the short-term and long-term.
(3 items) / Strongly disagree (1) to Strongly agree (5)
CISd (α = .80) / The Apollo moon landings never happened and were staged in a Hollywood film studio.
(4 items) / Strongly disagree (1) to Strongly agree (5)
STSe (α = .82) / The Earth, including all its inhabitants, is a living system.
(15 items) / Strongly disagree (1) to Strongly agree (5)
Religiosityf (α = .90) / How important are your religious or spiritual beliefs for what you do every day?
(5 items) / Not at all important (1) to Extremely important (5)

aRevised New Ecological Paradigm(Dunlap, van Liere, Mertig, & Jones 2000)

bConnectedness to Nature Scale(Mayer & Frantz 2004)

cFree Market Ideation(Heath & Gifford 2006)

dConspiratorial Ideation Scale(Lewandowsky, Gignac, & Oberauer 2013)

eSystems Thinking Scale(Davis & Stroink 2016)

fReligiosity(Strawbridge, Sherma, Cohen, Roberts, & Kaplan 1998)

Finally, participants were asked their age, gender, educational background, political affiliation (as Democrat, Independent, Republican, or Other), and political ideology (on a scale from 0, very liberal, to 100, very conservative).

In total, the survey took 25 minutes for participants to complete. In exchange, participants were given $2.50.

2.2 Reliability and Item Analysis

To ensure that the ratings of the metaphors were reliable, we computed Cronbach’s alpha for each of the three components of the scale. As noted in the main text, each of the components of the scale was internally consistent: Cronbach’s alpha was .74, .78, .76 for the questions related to the affect of the metaphors, the systemic-ness of the metaphors, and the accuracy of the metaphors, respectively.

We also computed the intraclass correlation of the ratings. This analysis is designed to test whether the metaphors were scored similarly across participants (i.e. did different people give similar ratings to, e.g., the home metaphor?). A high score would indicate that participants agreed on the degree to which each metaphor expressed positive affect, the degree to which each metaphor conveyed the idea that the earth is a complex system, and the degree to which each metaphor was accurate. This is what we found: the intraclass correlation (type = “agreement’, model = “two-way”) was .997, .995, and .994 for the affective, systemic, and accuracy components of the ratings, respectively.

Finally, we conducted an item analysis for each of the three components of the scale (see Table S3). Note that the questions about how well the metaphors convey the idea that humans influence the earth and are affected by the earth have are the most representative items for the systemic dimension (i.e., they are the most correlated with the five-item measure of how systemic the metaphors are; and they are the most reliable items). This is why these two questions were included to represent the systemic dimension of the metaphors in Study 2.

Table S3. Results of item analysis of ratings data for Study 1.

Dimension / Question / Correlation of item with total score / Reliability
Affect / Like / .78 / 1.43
Thought / .85 / 1.31
Familiarity / .82 / 1.37
Systemic / Complexity / .61 / 1.03
Connected / .77 / 1.21
Influence / .77 / 1.32
Affected / .78 / 1.34
Dynamic / .74 / 1.19
Accurate / Accurate / .89 / 1.52
Misleading / .90 / 1.63

2.3 Testing for Fatigue

The number of ratings that participants were asked to make and the amount of time it took participants to evaluate each metaphor (median = 44 seconds) is in line with several studies that have sought to gauge various linguistic properties of metaphorical sentences (e.g., Cardillo, Schmidt,Kranjec,Chatterjee, 2010; Cardillo, Watson, & Chatterjee, 2016; Katz, Paivio, Marschark, & Clark, 1988; Roncero & de Almeida, 2015). In addition, there was substantial variability in participants’ ratings (with, e.g., the home metaphor scoring 2.99 units, 1.80 units, and 1.69 units higher than the bank metaphor on the affective, systemic, and accuracy dimensions, respectively), suggesting that participants were engaged in the ratings task.

Nevertheless, we were concerned about the possibility that some participants may have become fatigued over the course of the experiment. For this reason, we have included two additional analyses below. In the first, we only analyze the first metaphor that each participant considered. In the second, we only analyze metaphors that participants spent at least 30 seconds evaluating.

2.3.1 Analyzing the First Metaphor Participants Evaluated

Each of the 17 metaphors was evaluated first by at least 34 participants (i.e. was the first metaphor rated by at least 34 participants). On this analysis (see Table S4), the metaphors elicited similar ratings as they did overall: there was a high correlation between ratings of affectiveproperties of the metaphors (when rated first compared to when all participants rated all of the metaphors), r[15] = .96, p < .001, properties related to systems thinking, r[15] = .84, p < .001, and the accuracy of the metaphors, r[15] = .90, p < .001.

Note, for example, that the home metaphor was rated as eliciting the most positive affect and as the most accurate on both analyses. It was rated as the most systemic when all participants rated all of the metaphors and among the most systemic when it was the first metaphor rated by participants. At the other end of the spectrum, the bank metaphor was rated as eliciting among the least positive affect, as only weakly expressing the idea that the earth is a complex system, and as one of the most misleading metaphors on both analyses.

The subtle differences between the means may result from participants comparing across metaphors as they progressed through the survey. For instance, a participant who rated the home metaphor first and the bank metaphor second likely considered questions about the affect, systemic-ness, and accuracy of the bank metaphor relative to the home metaphor.

Table S4. Mean ratings of metaphors when rated first and when evaluated by all participants in terms of affective andsystemic properties and accuracy.

Affect / Systemic / Accurate
First / Overall / First / Overall / First / Overall
Home / 5.39 / 5.36 / 4.88 / 5.12 / 6.14 / 6.17
Mother / 4.85 / 4.87 / 4.96 / 4.94 / 4.88 / 5.08
Community / 4.77 / 4.35 / 5.01 / 4.99 / 5.50 / 5.35
Network / 3.83 / 3.71 / 5.10 / 4.78 / 5.43 / 4.90
Investment / 4.44 / 3.88 / 5.29 / 4.74 / 5.71 / 4.93
Child / 3.75 / 3.57 / 4.62 / 4.64 / 4.35 / 4.24
Parent / 4.28 / 3.65 / 5.06 / 4.48 / 4.82 / 4.44
Farm / 3.94 / 3.53 / 4.65 / 4.50 / 4.94 / 4.79
Gift / 4.77 / 4.38 / 4.74 / 4.48 / 5.41 / 5.05
Ancestor / 3.93 / 3.49 / 4.77 / 4.30 / 4.81 / 4.34
Kingdom / 4.31 / 3.77 / 4.66 / 4.22 / 4.59 / 4.13
Machine / 3.74 / 3.04 / 4.12 / 3.90 / 4.26 / 3.72
Market / 3.72 / 2.87 / 4.54 / 3.86 / 4.87 / 3.77
Store / 3.11 / 2.47 / 4.21 / 3.38 / 3.94 / 3.27
Park / 3.72 / 3.02 / 4.19 / 3.71 / 4.59 / 4.01
Spaceship / 3.73 / 2.95 / 4.12 / 3.60 / 4.04 / 3.43
Bank / 3.29 / 2.37 / 4.37 / 3.32 / 4.45 / 3.15

2.3.2 Excluding Fast Responses

We also tested whether the ratings would differ if “fast” responses were eliminated. Figure S1 shows a histogram of the amount of time that participants spent evaluating each metaphor (median = 44 seconds). Participants who responded to the ratings questions extremely quickly (we have set a threshold of 30 seconds) may not have been as engaged in the ratings task as participants who spent a comparatively longer time thinking about each metaphor.

Figure S1. Histogram of the amount of time participants spent evaluating each metaphor (i.e. answering the 10 ratings questions and the question about who the likely source of the metaphor would be).

Overall, 21% of the metaphors were rated in 30 seconds or less. People spent a similar amount of time evaluating the 17 metaphors, and there was no evidence thatspecific metaphors were more likely to be evaluated in less than 30 seconds, χ2(16) = 4.72, p = .997. Mean ratings of the metaphors were nearly identical on the two analyses (i.e. when all data were included versus when only data from metaphors rated in longer than 30 seconds were included): with correlations of r[15] = .999, .996, and .999 for the affective, systemic, and accuracy ratings of the metaphors, respectively.

Based on the results of these two analyses, we do not find reason to believe that, e.g., the home metaphor elicited higher ratings on the three dimensions and, e.g., the bank metaphor elicited lower ratings on the three dimensions because participants became fatigued over the course of the experiment.

2.4 Individual Differences

Finally, we provide additional information about how well the metaphors resonated with subgroups from the general population. We focus on the six metaphors that were tested with both samples: home, mother, community, park, spaceship, and bank. We test for relationships between how people evaluated these six metaphors and seven individual difference measures: gender, political affiliation, age, religiosity, free market ideation, conspiratorial ideation, and belief in global warming. For brevity and parsimony, we created a single composite dependent measure that reflected the affect, systemic-ness, and accuracy of the metaphors for these analyses.

2.4.1 Gender

First we investigate whether males and females evaluated the metaphors differently. There were 330 males and 532 females in our sample, as well as five people who identified as neither male nor female (data from these five participants were excluded from the current analysis). Overall, males and females gave the highest ratings of the earth-as-home metaphor: with males, t[329]= 7.86, p .001,and females, t[531]= 10.43, p .001, rating the home metaphor more positively than the earth-as-mother metaphor (the second highest-rated metaphor for both groups). Females rated the home, mother, and communitymetaphors more positively than males; males rated the spaceship and bank metaphors more positively than females. See Table S5 and Figure S2.

Table S5. Mean ratings by gender. Statistically significant gender differences noted with asterisks: *p.05, **p<.01, ***p<.001.

Home** / Mother* / Community* / Park / Spaceship*** / Bank**
Male / 5.30 / 4.85 / 4.78 / 3.63 / 3.67 / 3.15
Female / 5.48 / 5.01 / 4.93 / 3.53 / 3.19 / 2.92

Figure S2. Left: pie chart showing breakdown of sample by gender; right: evaluations of six target metaphors by gender. Error bars denote standard errors of the means.

2.4.2 Political Affiliation

We then investigated whether the metaphors were evaluated differently as a function of participants’ political affiliation. In our sample, 347 people identified as Democrats, 293 as Independents, and 170 as Republicans. An additional 57 participants identified as Other and were not included in the current analysis.

The earth-as-home metaphor received the highest ratings from participants, regardless of their political affiliation. Democrats, t[346] = 5.97, p < .001,and Independents, t[292] = 7.53, p < .001, preferred the home metaphor significantly more than the mother metaphor (their second most highly evaluated metaphor); Republicans preferred the home metaphor to the earth-as-community metaphor, t[169] = 8.95, p < .001, which was the second highest rated metaphoramong Republicans (see Figure S3). There were no differences in how the groups rated the home, park, and bank metaphors; there were differences in how the groups rated the mother, community, and spaceship metaphors. Democrats and Independents tended to rate all three of these metaphors more highly than Republicans (see Table S6).