A good theory, the adage goes, should fit the data like a glove – covering the fingers (i.e., where the phenomenon is found) but not the space in between (where it is not; see Roberts and Pashler, Reference Roberts and Pashler2000 for a more formal treatment). How does the theory of grounded separation fare under this criterion?
First, is there a robust empirical phenomenon about which to theorize? Lee and Schwarz (L&S) concede that the basic phenomenon (e.g., less influence of past experience on decisions after hand-washing) is still under debate, while Ropovik, Sparacio, and Ijzerman (this treatment) and Ross et al. (this treatment) make compelling cases that the replicability of the basic phenomenon has not yet been established. A replication package for a strong unconfounded test of the phenomenon is needed, including defining any necessary pre-conditions in advance. This would enable a skeptical scientific field to either fully establish the robustness of the effect to its own satisfaction or demonstrate a lack of robustness in a way that would cause proponents to reconsider.
Should robustness be established, the question shifts to generalizability. Is the underlying phenomenon manifested sufficiently broadly across contexts and domains to require a general theory? The authors argue convincingly that the cleansing effects reported in the literature cannot all be explained in terms of either disgust reactions or cleansing as a morality metaphor. The proposed theory can indeed explain a broader range of potentially related effects. However, further research would be needed to determine whether phenomena ostensibly related to notions of connection or separation do in fact operate via the same psychological mechanism. It is not at all self-evident that any effects of closing up or of keeping stimuli, of prior ownership or of shopping via tablet computer on decision-making, for example, involve the same psychological process as cleansing.
The final challenge for an effective theory is to not only anticipate the presence of a phenomenon where it occurs, but also accurately predict absence of the phenomenon, where it does not. The grounded separation theory fails to specify when such effects would not occur, other than to propose modality as a moderator, because of differences in engagement of sensorimotor capacities (e.g., physical experience having stronger effects than conceptual activation). In fact, the authors are admirably precise about the intended expansiveness of the theory, stating that the theory would be falsified if “acts of separation, such as cleansing, do not result in any attenuation or elimination of an otherwise observed influence.”
The theory would therefore be complete if acts of separation (which are highly frequent, as the authors note) do nearly always attenuate or eliminate prior observed influences. The proposed theory therefore makes the remarkable claim that the norm in human psychology must be little or no influence of even the recent past on current attitudes, decision, and behaviors. Past influence should only be observed in the special case when typically ever-present separation-symbolic behaviors, such as handwashing, are absent.
This prediction directly contradicts research in a wide variety of domains, which has identified exactly the kind of long-term stability in attitudes, preferences, and behavior that should be “washed away” by people's frequent separation-symbolic actions. A large research literature has found that personality traits (e.g., the Big Five) are stable over multiple years, and the stability tends to increase over the life span (summarized in Roberts & DelVecchio, Reference Roberts and DelVecchio2000), with similar findings for religious belief and practice (Hamberg, Reference Hamberg1991). Similar stability over time has been documented for people's optimism (Billingsley, Waehler, & Hardin, Reference Billingsley, Waehler and Hardin1993), political party identification, and ideological orientation (Freeze & Montgomery, Reference Freeze and Montgomery2016; Green & Palmquist, Reference Green and Palmquist1994; Krosnick, Reference Krosnick1991). As a particularly striking example, given cleansing effects on optimism (Körner & Strack, Reference Körner and Strack2019), optimism and pessimism were largely stable over the course of a year among women undergoing cancer surgery, regardless of whether they received good or bad news about their condition (Schou, Ekeberg, Sandvik, & Ruland, Reference Schou, Ekeberg, Sandvik and Ruland2005).
Another research literature has studied state dependence in people's behavior, investigating whether stable patterns of behavior occur specifically because people's current choices are influenced by their past choices (as opposed to stability because of heterogeneous causal factors remaining the same). Evidence for state dependence, a persistent causal influence of prior choices (often years before) on subsequent behavior, has been found for moral behaviors (charitable giving and volunteering, Choi & Chou, Reference Choi and Chou2010; Meer, Reference Meer2013; criminal offending, Blokland & Nieuwbeerta, Reference Blokland and Nieuwbeerta2010; Nagin & Paternoster, Reference Nagin and Paternoster2000), morality-relevant behaviors (voting, Denny & Doyle, Reference Choi and Chou2009) and behaviors that are largely non-moral (consumer purchasing, Dubé, Hitsch, & Rossi, Reference Dubé, Hitsch and Rossi2010).
For example, consider a typical person's morning routine: she wakes up, showers, changes into her work clothes, closes her lunch into a container, exits her house, closes and locks the front door, dumps the kitchen garbage bag in the outdoor container, gets in her car and drives to work (trying to avoid other cars), and walks into her office. Between waking up and 9 AM, she has cleansed, changed, enclosed, destroyed, avoided, distanced, and changed context: all “grounded procedures” carried out through physical experience, ostensibly the most impactful modality.
Nevertheless, voluminous research and everyday experience both tell us that her personality, political views, morality, religious beliefs, tastes, and preferences will be quite unchanged, not only from the day before, but often even from the year before. She will generally engage in the same hobbies, chat with the same friends, support the same political party, donate to the same charities, shop at the same store for largely the same goods, and have similar feelings about her life and her future, with her past behavior continuing to drive her current behavior, no matter how many everyday acts of symbolic separations she conducts.
In short, the theoretical advance here is broadening without “tightening,” re-categorizing psychological phenomena without providing a more precise glove-like description of human behavior. The result, I fear, is instead a quite large mitten-theory with perhaps only a very small empirical hand inside; a theory of instabilities spotted in the lab that fails to account for the actual pervasive stability in the typical person's everyday life.
You have
Access
A good theory, the adage goes, should fit the data like a glove – covering the fingers (i.e., where the phenomenon is found) but not the space in between (where it is not; see Roberts and Pashler, Reference Roberts and Pashler2000 for a more formal treatment). How does the theory of grounded separation fare under this criterion?
First, is there a robust empirical phenomenon about which to theorize? Lee and Schwarz (L&S) concede that the basic phenomenon (e.g., less influence of past experience on decisions after hand-washing) is still under debate, while Ropovik, Sparacio, and Ijzerman (this treatment) and Ross et al. (this treatment) make compelling cases that the replicability of the basic phenomenon has not yet been established. A replication package for a strong unconfounded test of the phenomenon is needed, including defining any necessary pre-conditions in advance. This would enable a skeptical scientific field to either fully establish the robustness of the effect to its own satisfaction or demonstrate a lack of robustness in a way that would cause proponents to reconsider.
Should robustness be established, the question shifts to generalizability. Is the underlying phenomenon manifested sufficiently broadly across contexts and domains to require a general theory? The authors argue convincingly that the cleansing effects reported in the literature cannot all be explained in terms of either disgust reactions or cleansing as a morality metaphor. The proposed theory can indeed explain a broader range of potentially related effects. However, further research would be needed to determine whether phenomena ostensibly related to notions of connection or separation do in fact operate via the same psychological mechanism. It is not at all self-evident that any effects of closing up or of keeping stimuli, of prior ownership or of shopping via tablet computer on decision-making, for example, involve the same psychological process as cleansing.
The final challenge for an effective theory is to not only anticipate the presence of a phenomenon where it occurs, but also accurately predict absence of the phenomenon, where it does not. The grounded separation theory fails to specify when such effects would not occur, other than to propose modality as a moderator, because of differences in engagement of sensorimotor capacities (e.g., physical experience having stronger effects than conceptual activation). In fact, the authors are admirably precise about the intended expansiveness of the theory, stating that the theory would be falsified if “acts of separation, such as cleansing, do not result in any attenuation or elimination of an otherwise observed influence.”
The theory would therefore be complete if acts of separation (which are highly frequent, as the authors note) do nearly always attenuate or eliminate prior observed influences. The proposed theory therefore makes the remarkable claim that the norm in human psychology must be little or no influence of even the recent past on current attitudes, decision, and behaviors. Past influence should only be observed in the special case when typically ever-present separation-symbolic behaviors, such as handwashing, are absent.
This prediction directly contradicts research in a wide variety of domains, which has identified exactly the kind of long-term stability in attitudes, preferences, and behavior that should be “washed away” by people's frequent separation-symbolic actions. A large research literature has found that personality traits (e.g., the Big Five) are stable over multiple years, and the stability tends to increase over the life span (summarized in Roberts & DelVecchio, Reference Roberts and DelVecchio2000), with similar findings for religious belief and practice (Hamberg, Reference Hamberg1991). Similar stability over time has been documented for people's optimism (Billingsley, Waehler, & Hardin, Reference Billingsley, Waehler and Hardin1993), political party identification, and ideological orientation (Freeze & Montgomery, Reference Freeze and Montgomery2016; Green & Palmquist, Reference Green and Palmquist1994; Krosnick, Reference Krosnick1991). As a particularly striking example, given cleansing effects on optimism (Körner & Strack, Reference Körner and Strack2019), optimism and pessimism were largely stable over the course of a year among women undergoing cancer surgery, regardless of whether they received good or bad news about their condition (Schou, Ekeberg, Sandvik, & Ruland, Reference Schou, Ekeberg, Sandvik and Ruland2005).
Another research literature has studied state dependence in people's behavior, investigating whether stable patterns of behavior occur specifically because people's current choices are influenced by their past choices (as opposed to stability because of heterogeneous causal factors remaining the same). Evidence for state dependence, a persistent causal influence of prior choices (often years before) on subsequent behavior, has been found for moral behaviors (charitable giving and volunteering, Choi & Chou, Reference Choi and Chou2010; Meer, Reference Meer2013; criminal offending, Blokland & Nieuwbeerta, Reference Blokland and Nieuwbeerta2010; Nagin & Paternoster, Reference Nagin and Paternoster2000), morality-relevant behaviors (voting, Denny & Doyle, Reference Choi and Chou2009) and behaviors that are largely non-moral (consumer purchasing, Dubé, Hitsch, & Rossi, Reference Dubé, Hitsch and Rossi2010).
For example, consider a typical person's morning routine: she wakes up, showers, changes into her work clothes, closes her lunch into a container, exits her house, closes and locks the front door, dumps the kitchen garbage bag in the outdoor container, gets in her car and drives to work (trying to avoid other cars), and walks into her office. Between waking up and 9 AM, she has cleansed, changed, enclosed, destroyed, avoided, distanced, and changed context: all “grounded procedures” carried out through physical experience, ostensibly the most impactful modality.
Nevertheless, voluminous research and everyday experience both tell us that her personality, political views, morality, religious beliefs, tastes, and preferences will be quite unchanged, not only from the day before, but often even from the year before. She will generally engage in the same hobbies, chat with the same friends, support the same political party, donate to the same charities, shop at the same store for largely the same goods, and have similar feelings about her life and her future, with her past behavior continuing to drive her current behavior, no matter how many everyday acts of symbolic separations she conducts.
In short, the theoretical advance here is broadening without “tightening,” re-categorizing psychological phenomena without providing a more precise glove-like description of human behavior. The result, I fear, is instead a quite large mitten-theory with perhaps only a very small empirical hand inside; a theory of instabilities spotted in the lab that fails to account for the actual pervasive stability in the typical person's everyday life.
Conflict of interest
None.