Investigating whether speakers use different linguistic means to report memories of personally experienced events and memories of general facts.
Paradigmatic cases of episodic recall differ from recall of semantic information by having special phenomenological properties (e.g., experience-likeness, mental imagery, perspectivity, sense of self). Recent work in philosophy and linguistics suggests that languages (here: the languages of the research unit, viz. German and English) have a wealth of markers for these properties. These include special adverbs like ‘vividly’ (for mental imagery), pronouns like German ‘man’ (for 3rd person perspective), and grammatical structures like ‘remember V-ing’ (for 1st person perspective). Our project uses these linguistic markers for episodic recall to test specific assumptions of the scenario construction framework (Cheng, Werning, & Suddendorf, 2016). This holds esp. for the informational sparseness of episodic memory traces [see P0’s Aim 4], the supplementation of trace information with semantic information during retrieval [Aim 2], and the influence of the self-model on scenario construction [Aim 3]. To test these assumptions, we will study the effect that episodicity markers have on the suitability of a given memory report (i.e., sentences of the form ‘I remember ...’) for the description of a certain remembering event (Obj1). We argue that, to capture the effect of some of these markers (Obj2), one must adopt a generative account that centrally involves scenario construction (Obj3). To better understand the effect of episodicity markers (see Obj1), we will conduct a series of online behavioral studies (WP1-3). In these studies, participants are asked to identify one out of two agents who is more likely to have uttered a certain ‘remember’-sentence (given their previous experience) resp. to choose one of two memory reports that best describes a given remembering event. These studies are necessary since existing work on episodicity markers has focused only on corpus and production data, and has not provided information about the effect that these markers have on the truth of the memory reports containing them. To account for these markers' effect (see Obj2), we will specify the (compositional, truth-conditional) meaning of each marker and describe how this meaning interacts with the meaning of the verb ‘remember’ to yield the truth-conditions from WP1-3 (WP4). This account will build on our earlier work on single-type semantics. This work centrally involves informationally rich, perspectival content and assumes the construction of scenarios from an informationally minimal initializing situation. Our previous work leads us to expect that the effect of some of these markers (e.g., the perspective-shifting effect of ‘man’) can only be explained by accounts that assume key features (esp. sparse traces, semantic supplementation, and retrieval bias) of the scenario construction framework (WP5). We will use our findings about episodicity marking to help elicit and interpret free recall data from the experimental projects of the research unit (Obj4, WP6).