Table 1
Explanations and Examples for the Employed Technical Terms.
| TERM | EXPLANATION | EXAMPLE |
|---|---|---|
| Amodal | A representation (e.g., of meaning) that does not include activity in sensory, motor, or emotional cortices or bodily activity such as gestures. | Abstract concepts, such as liberty, are frequently conceptualized as representations devoid of sensorimotor aspects (however, see Borghi et al., 2019). |
| Distributional semantic models | A class of computational models based on the hypothesis that words with similar meanings have similar distributions over linguistic contexts. These distributions can be extracted from corpora of natural language. | The concepts whale and dolphin occur in highly similar linguistic contexts (with sea, fish, ocean as frequent words in close vicinity) and can therefore be assumed to have similar meanings. |
| Embodied (see also Grounded and Situated) | The second level in Fischer’s (2012) proposed hierarchy of knowledge representation. Sensorimotor associations resulting from previous experiences, such as actions and their outcomes | Right-handed people associate the right side in space with good and the left side in space with bad, whereas left-handed people show the opposite association (Casasanto, 2009). |
| Grounded (see also Embodied and Situated) | The first level in the hierarchy of Fischer (2012). Associations whose origins lie in the way the physical world is structured. | Accumulating objects cannot permeate one another, leading to associations of more with up (Lakoff & Johnson, 1980). |
| Iconicity | Non-arbitrary relations between sublexical elements of (spoken or signed) words and denoted concepts. | Association between high-frequency vowels (e.g., /i/ vs. /a/) and small (vs. large) size. |
| Imagine manipulation (see also Physical manipulation) | The second process taught to children (in Glenberg et al., 2004) to help them index words, phrases, and syntax to their meanings. When reading texts, children are asked to imagine moving images or toys into a configuration that accurately portrays the meaning of the text. | When reading that a farmer brings the cart to the barn, children would imagine the toy farmer moving to the toy cart and then both moving to the toy barn. |
| Indexing | The process of mapping a linguistic term onto its referent. | Mapping the word horse onto a toy horse. |
| Physical manipulation (see also Imagine manipulation) | The first process taught to children (in Glenberg et al., 2004) to help them index words, phrases, and syntax to their meanings. When reading texts, children are asked to physically move images or toys into a configuration that accurately portrays the meaning of the text. | When reading that a farmer brings the cart to the barn, children would move the toy farmer to the toy cart and then move both to the toy barn. |
| Sensorimotor simulation | In the context of language comprehension, sensorimotor simulation is activity in sensorimotor and emotional cortices in response to linguistic stimuli that is homologous to the activity engendered by the perception of the described objects, events, or activities. | When reading the word salty, gustatory brain regions would be active in a way similar to when tasting something salty. |
| Situated (see also Grounded and Embodied) | The third level in Fischer’s (2012) hierarchy. The human cognitive system is influenced by its current environment, adapting in a flexible manner to current goals and task constraints even when long-standing embodied or grounded knowledge conflicts. | One’s current state of high (vs. low) fatigue lead to hills being estimated as more (vs. less) steep (Bhalla & Proffitt, 1999). Depending on the number range (0–5 vs. 4–9), the numbers 4 and 5 are associated with either right or left space (Dehaene, Bossini, & Giraux, 1993). |
| Symbol grounding-by-language | Language can be conceptualized as experience-by-proxy (Johnson-Laird, 1983) or second-hand experience. Symbol grounding-by-language describes the idea that not only primary sensorimotor experience, but also this secondary experience can establish grounding for symbols. | For someone who has no direct sensorimotor experience with the concept zebra, reading that a zebra is a horse with black and white stripes will provide grounding for zebra. Thus, language alone will lead to a representation of the concept that incorporates sensorimotor information (Harnad, 1990). |
| Symbol grounding problem | The observation that language cannot be a self-contained system, in which each symbol is only defined by other symbols, as this would constitute an infinite regress of dependencies (Harnad, 1990). | Trying, as a thought experiment, to learn Chinese using nothing but a Chinese/Chinese dictionary would mean seeing unknown symbols when looking up an unknown symbol, leading to an infinite pass from symbol to symbol without gaining any understanding of Chinese (Harnad, 1990). |
Figure 1
Mean reaction times for 9 conditions (combinations of semantic modalities: Word 1 × Word 2; see main text for details). Whiskers represent standard errors. Orange horizontal lines represent significant differences between conditions (p = .02 or lower).
Figure 2
Data from Günther, Nguyen et al. (2020). Distribution over participants (lighter left part) and items (darker right part). The solid lines indicate the mean values. Reprinted from the Journal of Memory and Language, Volume 115, Günther, Nguyen, Chen, Dudschig, Kaup, & Glenberg, Immediate sensorimotor grounding of novel concepts learned from language alone, 2020, with permission from Elsevier.
Figure 3
Data from Morey et al. (2021). Within each lab, the middle horizontal line indicates the median effect, and the two other lines indicate the inter-quartile range. Points are translucent, meaning that darker areas indicate overlapping points and thus higher density. Reprinted under the Creative Commons license (CC BY 4.0).