In linguistics and moral psychology, the "AGENT" is the doer of an event; the "PATIENT" is the person affected by the event: the person who has the thing done to them. We can think of ourselves and others as agents and patients of moral and emotional experiences.

In this schematic of an event below, say Amy is the AGENT, and Bill is the PATIENT. If Amy hit Bill, it seems reasonable that Amy would be considered "causal" since she had to muster up the energy and strike the blow. 

lauraniemi.com
Source: lauraniemi.com

But what if Amy liked Bill? 

Or, if Amy abandoned Bill?

Who caused these?

And who's responsible?

The causal information that comes "for free"—already encoded in verbs—has been investigated in linguistics. Intriguingly, there are reliable patterns in people's causal judgments about some events (Bott & Solstad, 2014; Hartshorne & Snedeker, 2013). Some verbs just compel us to expect AGENT causes more than PATIENT causes, and vice versa.

For example, verbs related to judgment (e.g., praise, punish, congratulate; (Kipper-Schuler, 2006, VerbNet) cluster together and reliably compel us to expect that the affected PATIENT caused them.

When "Amy punished Bill," we presume that Bill did something that caused Amy to punish him (Bott & Solstad, 2014).

Other verbs, including verbs in a class that includes the verb "harm", tend to compel us to expect that the AGENT caused them, such as delight, confuse, irritate.

When "Amy irritated Bill," we presume that something Amy did caused Bill to be irritated.

Other verbs presented in these simple frames, including many conveying morally-relevant events like killing, coercing, and clobbering, don't necessarily compel expectations about AGENT or PATIENT causes. This is where people's moral values and motivations come in.

We've found that people's higher-level moral values—like how much they value universal human rights versus how much they value loyalty to their special group and obedience to authority—factor into their expectations about the causes. People who highly value loyalty and obedience, for example, are more likely to select the affected PATIENT as causal (Niemi, Hartshorne, Gerstenberg & Young, 2016).

 Western History/Genealogy Department, Denver Public Library "Homosexuals being punished by wearing dresses and wheeling heavy rocks." State Penitentiary in Canon City, Colorado.
Source: Wikimedia Commons: Western History/Genealogy Department, Denver Public Library "Homosexuals being punished by wearing dresses and wheeling heavy rocks." State Penitentiary in Canon City, Colorado.

This aligns with other findings that people with high loyalty and obedience values also explicitly blame victims more (Niemi & Young, 2016) and say they "deserved" and "allowed" the bad events (Niemi et al., 2016). These findings suggest that judgments about the causes of events like killing and coercing may hinge on whether these events are perceived as potentially relevant to (patient-caused) punishment or (agent-caused) harm, which may be tied to the moral values of loyalty and obedience to authority.  

Also, we found that people's attitudes about gender, like whether they hold hostile attitudes about women, predicted their expectations about the likelihood that women cause men to harm women.

For example, given:

Bill stabbed Amy because... he or she

...people higher in hostile sexism (Glick & Fiske, 1997) were much more likely to select "she" than people low in hostile sexism (Niemi et al., 2016). 

We may prefer to think our moral values or social biases are in check when we reason about the goings-on of random strangers. Surely, there is a lot more work to be done to understand how much our moral commitments factor into our judgments of generic events. The evidence so far, however, indicates that our moral worldviews pervade our quickest inferences about interpersonal events. Our cognitive models of events seem closer to little moral dramas with toy agents and patients acting according to stereotyped scripts about typical causes, rather than stripped down schematics like the diagram above (Pinker & Levin, 1991).

It is for future research to understand how changing causal models might change explicit attitudes and behaviors, and vice versa. 

References

Hartshorne, J. K., & Snedeker, J. (2012). Verb argument structure predicts implicit causality: The advantages of finer-grained semantics. Language and Cognitive Processes, 28(10), 1474–1508. 

Bott, O., & Solstad, T. (2014). Psycholinguistic approaches to meaning and understanding across languages (studies in theoretical psycholinguistics). In B. Hemforth, B. Mertins, & C. Fabricius-Hansen (Eds.), (p. 213-251). Cham: Springer. 

Garvey, C., & Caramazza, A. (1974). Implicit causality in verbs. Linguistic Inquiry, 5(3), 459-464. 

You are reading

Morality in Language

What Counts as "Fair" and What Makes People Care?

The neural underpinnings of "fairness" as impartiality, charity and reciprocity

Linguistic Smoke and Mirrors in Leadership

Speaking vaguely has special implications for political leaders.

Who's Responsible for My Emotional Experience?

Our cognitive models of events can be like little moral dramas.