Do people respond to machines mindlessly or reflectively?

Van der Goot and Etzrodt argue that human-machine communication (HMC) research has run on two fundamentally different, usually unacknowledged, explanatory paradigms. The Media Equation paradigm, originating in Reeves and Nass's 1996 claim that "media equal real life," holds that people respond mindlessly to machines "as if they are humans" even while knowing better. The Media Evocation paradigm, drawing on Turkle's work on evocative objects (1984, 2007), instead treats machines as objects "betwixt and between" categories like person and thing, which provoke conscious reflection rather than automatic response. The authors point to the CASA acronym itself as evidence the field has blurred the two: "computers as social actors" (Media Equation's as-if stance) and "computers are social actors" (Media Evocation's ontological stance) get used interchangeably, "usually without addressing this difference."

The mechanisms the two paradigms give are not variants of each other but different theories of why social responses happen at all. Media Equation's mechanism is cue-triggered and automatic: Reeves and Nass write that responses can be initiated by "minimal cues," because "modern media engage old brains" not evolved for 20th-century technology — "absent a significant warning that we've been fooled, our old brains hold sway." Media Evocation's mechanism is categorical and reflective: following Turner (1969) on liminal states and Douglas (1966) on marginality, Turkle frames machines as possessing neither the properties of what they used to be ("thing") nor of what they might become ("subject" or "hybrid"), which forces users into negotiation about the machine, themselves, and their relationship — "computers bring philosophy down to earth" by raising questions about the machine's "mind" that extend to questions about our own.

This sharpens rather than duplicates several nearby notes. Do humans apply human-human scripts to AI interactions? refines the Media Equation side specifically — it keeps the mindless-script mechanism but corrects what the scripts are borrowed from (media-specific, not human-human), without moving toward Evocation's ontological framing. Do more social cues always make AI feel more present? already names both mindless and mindful anthropomorphism as drivers — that is the dual-process bridge (via Langer's mindlessness/mindfulness work) that van der Goot and Etzrodt call for explicitly, suggesting the MASA line is already operating across both paradigms rather than settling into one. Why do people share more openly with machines than humans? stays inside Media Equation's mechanistic idiom — suppressed goals from absent inner experience — rather than asking what kind of social actor a machine ontologically is.

The paper is theoretical, not a new empirical study: its one cited statistic — that 132 HMC publications across 28 communication journals over the past decade were 48.5% quantitative and 40.2% experimental — is background evidence that the field has operationalized Media Equation far more than Media Evocation, not a finding the authors generated themselves. The paper does not specify how to operationalize Evocation-style reflection empirically; it calls for mixed-method, qualitative-inclusive research as future work rather than supplying it. The implication is that a survey or experiment showing people treat AI like a person is evidence for the Equation paradigm only — it cannot confirm or rule out the separate, harder-to-measure Evocation-style ontological negotiation the same interaction might also be producing.