Running across the mind or across the park: does speech about physical and metaphorical motion go hand in hand?

1.1 Expression of physical motion events

Languages can be grouped into two main typological categories, depending on how they choose to encode path of motion (i.e., the trajectory of a moving entity)—a preference that has consequences for the lexicalization of manner (i.e., the movement style of a moving entity; Talmy 2000). S-languages, such as German, English, and Polish opt to lexicalize path in a satellite outside the verb (e.g., a prefix, a particle, a prepositional phrase), leaving the verb root free to encode manner (e.g., German: Ein Mann läuft ins Haus ‘A man runs into the house’).^[1] In contrast, V-languages such as Spanish, Turkish, and Japanese typically express path in the main verb and manner in an optional subordinate clause (e.g., Spanish: Un hombre entra en la casa [corriendo] ‘A man enters the house [running]’).

These typological differences thus result in different strategies in the packaging of motion components by adult native speakers: S-language speakers predominantly rely on conflated constructions in which path and manner are expressed within a single clause consisting of a main verb and a path satellite, whereas V-language speakers show a preference for separated constructions in which path and manner appear in two independent constituents (e.g., Kita and Özyürek 2003; Özçalışkan et al. 2016a, 2018)—a differential pattern that has been shown across several different languages of the world (Akhavan et al. 2017; Allen et al. 2007; Fritz et al. 2019; Lewandowski and Özçalışkan 2018, 2021; Tütüncü et al. 2023).

The preferred pattern of packaging also has consequences for the lexicalization of motion events at the level of both the main verb and secondary motion expressions outside the verb. The verb typically carries manner information in S-languages and path information in V-languages. Accordingly, S-language speakers use greater amount (i.e., verb tokens) and variety (i.e., verb types) of manner verbs than V-language speakers, while V-language speakers follow the opposite pattern in their descriptions of motion events (e.g., Cardini 2010; Hickmann et al. 2009; Ibarretxe-Antuñano 2009; Özçalışkan and Slobin 1999; Tusun and Hendriks 2019).

The encoding of manner versus path in the main verb also impacts the expression of the two components in secondary motion expressions outside the verb (e.g., particles, prepositions, manner adverbs). In S-languages, in which the main verb is mostly reserved for manner information, secondary motion elements become the place where path information is typically encoded (Beavers et al. 2010; Imbert 2012). As shown in earlier work, S-language speakers frequently use particles, prepositions, and prefixes to express path of motion outside the verb with multiple path segments within a single motion construction (Choi 2009; Emerson et al. 2021; Lewandowski and Mateu 2016; Özçalışkan 2015; e.g., He ran out of the house into the yard and then down to the river). In contrast, with path being lexicalized in the verb root, V-languages are generally required to encode each path segment in a separate clause with its own verb (e.g., Spanish: Él salió de la casa corriendo, pasó por el jardín y bajó al río ‘He exited the house running, passed through the yard, and descended to the river’). The greater syntactic complexity associated with using multiple main verbs makes V-language speakers more likely to express a single path segment per motion event—thus providing less specific descriptions of motion trajectories in V-languages than in S-languages (e.g., Lewandowski 2022; Özçalışkan 2015; Slobin 2004). As a corollary to this, in V-languages, given that the main verb is typically reserved for path of motion, secondary motion elements become the place where secondary manner information can be expressed. At the same time, however, existing research (e.g., Emerson et al. 2021; Özçalışkan 2009, 2015; Özçalışkan and Slobin 2003; Slobin 2004) suggests that V-language speakers rarely express manner in their secondary motion expressions, but instead use them primarily to convey additional path information.

Relatively less is known about more granular differences in the expression of physical motion between languages from the same typological affiliation (i.e., intra-typological variation). Existing work that examines how packaging patterns vary along intra-typological lines is limited to only one study, namely Soroli et al.’s review of existing work (Papafragou et al. 2008; Soroli 2011; Soroli and Verkerk 2017; in Soroli et al. 2019) that suggested differences in the packaging of physical motion events in Greek and French (both V-languages): French adult speakers showed closer alignment to the separated pattern than Greek speakers in their description of animated motion scenes. The few existing studies that focused on the more detailed aspects of lexicalization patterns (i.e., motion verbs, secondary motion expressions), however, suggests some differences, particularly in the extent with which languages encode manner and path information (e.g., Filipović 2007; Hijazo-Gascón and Ibarretxe-Antuñano 2013; Ibarretxe-Antuñano 2004, 2009; Lewandowski 2018, 2021; Morita 2011; Sugiyama 2005).

With regard to path of motion, Ibarretxe-Antuñano (2009) observed that languages can be placed on a continuum ranging from low-path-salient to high-path-salient languages—a cline that goes beyond the two-way split between V- and S-languages. For instance, although Spanish and Basque are both V-languages, speakers of Spanish typically resort to lexicalizing path in the main verb, while speakers of Basque commonly provide more extended path elaborations by encoding additional segments (e.g., source, goal) in secondary motion elements (Ibarretxe-Antuñano 2004; see Özçalışkan 2009, for a similar pattern in Turkish). Similarly, German versus Polish, both S-languages, show variability in the lexicalization of path in secondary motion elements (e.g., Lewandowski 2022; Lewandowski and Mateu 2016): German speakers are more likely to attach a greater number of path segments to a single verb of motion as compared to Polish speakers, thus suggesting intra-typological variability in the lexicalization of path.

With regards to manner of motion, researchers also argue that languages form a cline of manner salience, where some languages elaborate more granular manner distinctions in the main verb compared to others even within the bounds of a single typological group (Slobin 2004). For example, motion verbs in Germanic languages, such as English and German, express a greater range of manner variations (i.e., manner verb types) than Slavic languages (e.g., Polish, Serbian) (Filipović 2007; Kopecka 2010; Lewandowski 2018, 2021), showing intra-typological variability in the lexicalization of manner.

The observed intra-typological contrasts–both for path and manner of motion—are likely to be related to differences in the availability of morpho-syntactic and lexical devices for motion expression in languages of the same typological affiliation (Beavers et al. 2010). For example, Basque, unlike Spanish, has a rich repertoire of directional cases and postpositions, allowing its speakers to encode additional path elements outside the verb root more easily than Spanish speakers (Ibarretxe-Antuñano 2004). In a similar vein, Germanic languages typically lexicalize path in morphologically free elements (e.g., particles), while Slavic languages do so in morphologically bound elements, such as prefixes (e.g., Filipović 2010; Lewandowski 2021). The availability of morphologically free satellites provides an easier option to combine path satellites with a greater variety of manner verbs, thus differing from morphologically bound satellites which offer more limited options to combine with roots encoding manner. This, in turn, results in fewer types of manner verbs and lower incidence of secondary path expressions in Slavic languages compared to Germanic languages even if both groups are classified as S-languages (Lewandowski 2021; Lewandowski and Mateu 2016, 2020).

Overall, research to date suggests strong inter-typological differences in the encoding of physical motion for both packaging and lexicalization of motion elements. S-language speakers rely more on conflated strategies, with greater token and type frequencies of manner verbs and secondary motion expressions expressing path of motion. V-language speakers, on the other hand, opt to rely more on separated packaging strategies, with greater token and type frequencies of path verbs but relatively infrequent secondary motion expressions expressing manner of motion. The intra-typological variability becomes evident largely in the lexicalization of motion, particularly with respect to the extent (tokens) and diversity (types) with which path and manner components are expressed either in the verb or in a secondary motion element within a language type.

1.2 Expression of metaphorical motion events

Metaphors involve a conceptual mapping from a relatively more concrete source domain to an abstract target domain (Lakoff and Johnson 1980, 1999). Importantly, movement in space serves as a frequently used source domain in structuring a wide set of abstract concepts including ideas (e.g., This idea bounced back into my mind), emotions (e.g., He fell in love), time (e.g., The hours slip by), along with many others across different languages (e.g., Kövecses 2005; Lakoff and Johnson 1980; Moore 2006; Özçalışkan 2003). There is no existing work that has examined packaging patterns in the expression of metaphorical motion. However, most of the earlier research that focused on the lexicalization of metaphorical motion in typologically distinct languages suggests that patterns of crosslinguistic variability observed in the description of physical events largely apply to the descriptions of metaphorical events (Caballero and Ibarretxe-Antuñano 2015, Cifuentes-Férez 2014; Matsumoto 1996, Özçalışkan 2004, 2005; Wilson 2005).

In one of these earlier studies, Özçalışkan (2003, 2004, 2005 examined lexicalization patterns of metaphorical motion events in written texts (novels, newspapers) and oral responses produced by native English (S-language) and native Turkish (V-language) speakers. The patterns of metaphorical motion expressions in Özçalışkan’s studies largely mirrored the patterns observed in earlier work on physical motion (e.g., Özçalışkan and Slobin 1999; Slobin 1996). In their metaphorical motion descriptions, productions in English included greater amount and variety of manner verbs than the ones in Turkish—a pattern that was reversed for path verbs. The differences were also evident in secondary motion expressions: in both languages, written and oral productions occasionally included manner expressions outside the verb, mainly to further elaborate the manner information that was already lexicalized in the main verb in English and to add new manner information that was not expressed in the main verb in Turkish. The lexicalization of path outside the verb was common in both languages, but it was more frequent in the productions in English compared to the ones in Turkish (Özçalışkan 2004, 2005). These patterns were also evident in other studies that examined metaphorical motion events in other V-languages, including Greek (Katis and Selimis 2005), Japanese (Matsumoto 1996), and Spanish (Caballero and Ibarretxe-Antuñano 2015; Cifuentes-Férez 2014; Wilson 2005). All of these earlier studies on inter-typological variation in metaphorical motion focused on lexicalization patterns—mostly at the level of the main verb—and provided further evidence for higher production of path verbs in V-languages compared to S-languages (see Özçalışkan et al. 2017 for a recent review).

However, there are also a few other studies that suggested that expression of metaphorical motion might differ from that of physical motion. An earlier analysis of oral and written descriptions of fictive motion (i.e., metaphorical conceptualization of perceptual information as a moving entity; e.g., the river runs through the valley; cf. Matlock 2004) showed that adult native speakers of English (S-language) showed similarities to Spanish speakers (V-language), using mostly path but rarely manner information in their responses, thus deviating from S-language patterns in describing metaphorical events (Rojo and Valenzuela 2003). A similar pattern of results was observed by Feist and Duffy (2020) in their corpus-based comparison of metaphorical motion descriptions in English and Spanish for the target domain of time: speakers of both languages frequently omitted manner, expressing primarily path in their metaphorical motion descriptions. These findings thus suggest variability in the expression of motion based on event type (i.e., greater exclusion of manner in metaphorical motion events across speakers of both S- and V-languages).

One possible explanation for this variability might be the nature of the two event types. Different from physical motion events, metaphorical events involve a mapping from the source domain of physical motion to the target domain of abstract concept (Lakoff and Johnson 1980, 1999). As such, the metaphorical event description is selective in the sense that it only highlights features of the source domain that are most relevant to the target domain, while downplaying other features that are not as relevant (e.g., Boers 1997; Koller 2022; Lakoff and Johnson 1980). Accordingly, the relative expression of manner and path components of motion, as well as their packaging, might show variability in the description of physical and metaphorical events within a particular language.

At the same time, the existing work on inter-typological variation in metaphorical motion remains relatively sparse, focusing only on a few V-languages all compared to one S-language (English). There is also no study that has yet compared expression of physical and metaphorical motion within a single study design across a broader set of abstract concepts (i.e., beyond fictive motion or time), further highlighting the need for future research in this domain.

Research on intra-typological variability in the expression of metaphorical motion is mostly non-existent. One earlier study (Zlatev et al. 2012) examined metaphorical motion descriptions of emotional states in languages that belonged to either the same (Bulgarian, English, Swedish, all S-languages) or different typological groups (Thai). This earlier study, however, was primarily a phenomenological analysis of the relationship between patterns of human experience and metaphorical expressions for the target domain of emotions and did not specifically focus on intra-typological variability. As such, analysis of intra-typological variability in the lexicalization and packaging of metaphorical motion still presents an important gap in the literature.

Overall, research to date suggests that expression of metaphorical motion in V-languages compared to English (as the S-language) closely aligns with the expression of physical motion, particularly in the lexicalization of motion—with the exception of a few studies that suggest differences for particular target domains. V-language speakers use greater amount and variety of path verbs, thus differing from S-language speakers who show more extended use of manner verbs combined with secondary motion expressions conveying path of motion. The paucity of research on intra-typological variation as well as packaging patterns in the expression of metaphorical motion, particularly across a broader set of S-languages, highlights these domains as important areas in need of future research.

1.3 Current study

Earlier research that focused either on physical or metaphorical motion events provides evidence of strong inter-typological differences in the packaging and lexicalization of motion, along with some intra-typological differences, particularly in the lexicalization of physical motion. In this study, we took these findings one step further by examining patterns of similarities and differences in the expression of both physical and metaphorical motion within the same data set. Our goal was to situate the extent of cross-linguistic variability in the two event types (physical, metaphorical), approaching it both from inter-typological versus intra-typological perspectives. More specifically, we examined physical and metaphorical motion descriptions produced in two relatively understudied S-languages—German and Polish—comparing them to Spanish, a V-language. German and Spanish are two paradigmatic instances of S- and V-languages, respectively, both in terms of the packaging and lexicalization of motion (e.g., Bamberg 1994; Cifuentes-Férez 2008; Harr 2012; Sebastián and Slobin 1994; Talmy 2000). Polish, in turn, differs from prototypical S-languages, particularly in its lexicalization of manner and path, with less specific encoding of manner in the main verb (i.e., fewer manner verb types) and less frequent expression of path in secondary motion elements (Kopecka 2010; Lewandowski 2021). As such, the languages under study would provide a highly relevant testing ground for the effects of both inter- and intra-typological factors in the expression of physical and metaphorical motion.

Our study used literary texts originally written in each of the three languages, with the goal to extend findings of earlier studies on oral elicitations to a different discourse genre of speech production, thus expanding the generalizability of the earlier findings. We used a two-pronged approach by focusing on both packaging strategies speakers employ in arranging manner and path components of motion (conflated vs. separated) and lexicalization strategies that become evident in the more detailed aspects of the motion event descriptions (choice of manner vs. path verbs, inclusion of secondary motion elements encoding manner or path). Our two-pronged approach (packaging and lexicalization) provides two different yet complementary angles on motion event encoding. Packaging provides an account of morpho-syntactic arrangement of path and manner at the constructional level; lexicalization, on the other hand, provides an account of more detailed manner and path expression at the level of more specific clausal components (i.e., verbal, nonverbal). This two-pronged approach also becomes particularly important when examining motion expression from inter- and intra-typological angles simultaneously, particularly given earlier work (e.g., Filipović 2007; Ibarretxe-Antuñano 2004) that suggests that more subtle intra-typological differences are more likely to emerge at the lexical, but not at the packaging level of analysis. As such, inclusion of both levels of analysis has the potential to provide a more comprehensive account of patterns of similarities and differences along inter-typological and intra-typological lines in the expression of physical and metaphorical motion. In this study, we asked two questions:

1. We first asked whether literary texts written in structurally different languages would show the same patterns of inter-typological and intra-typological variability previously observed in the spoken productions of adult native speakers in their expression of physical motion events. We predicted that writers would show strong inter-typological differences in the packaging and lexicalization of motion elements for physical motion, based on earlier work (e.g., Kita and Özyürek 2003; Özçalışkan et al. 2016a; Slobin 2004). We expected that German and Polish writers would use more conflated descriptions, along with greater amount and types of manner verbs and more secondary path expressions compared to Spanish writers. In contrast, we expected Spanish authors to use more separated descriptions along with greater amount and types of path verbs, compared to German and Polish writers. We expected no inter-typological differences in the use of secondary motion expressions encoding manner of motion, based on earlier work (Özçalışkan 2009; Slobin 2004).

   We also expected some intra-typological differences at the lexical level. We predicted that German authors would use greater variety of manner verbs, coupled with more descriptions with secondary path expressions than their Polish counterparts, consistent with earlier work (Filipović 2007; Kopecka 2010; Lewandowski 2021; Lewandowski and Mateu 2016; Lewandowski and Özçalışkan 2021).

2. We next asked whether the patterns that we observed in the expression of physical motion events would remain similar in the expression of metaphorical motion events in written texts. We predicted that the inter-typological differences in the packaging and lexicalization of metaphorical motion would be largely similar to the patterns in the expression of physical motion events, based on earlier work that showed close coupling between physical and metaphorical motion expression in languages from opposite typologies (i.e., Greek, Spanish or Turkish compared to English; Cifuentes-Férez 2014; Özçalışkan 2005). We also predicted that this overall similarity might be accompanied by more subtle and target domain-specific differences between the two event types, particularly with respect to the lexicalization of manner—a prediction that was based on earlier work (Feist and Duffy 2020; Rojo and Valenzuela 2003) that suggested lower expression of manner in metaphorical motion descriptions.

   Even though there is no existing work on intra-typological variability in metaphorical motion event descriptions, we predicted that the patterns of intra-typological variability evident in the lexicalization of physical motion would also become largely observable in the lexicalization of metaphorical motion, with greater variety of manner verbs and secondary path expressions in German compared to Polish. We also predicted that this might be accompanied by more subtle differences in the encoding of manner, with less frequent expression of manner in metaphorical descriptions produced by both Polish and German writers.

Overall, our study provides a unique account of physical versus metaphorical motion encoding across languages from the same (i.e., inter-typological variation) versus different (i.e., intra-typological variation) typological groups within a single study design—thus expanding our conceptual understanding of the cross-linguistic factors that underlie the linguistic construal of motion in space. Metaphorical motion constitutes a pervasive aspect of everyday communication for native learners. At the same time, it might present added challenges for non-native learners as the metaphorical mappings as well as their linguistic expression might differ across languages (Kövecses 2005). Accordingly, understanding patterns of variability in metaphorical expression between languages that are not only typologically different but also typologically similar might provide the basis on which more effective instructional strategies can be developed in second language learning contexts.

2.1 Sample

Our sample consisted of 45 contemporary fictional novels written originally in German, Polish, or Spanish, with 15 novels per language.^[2] We selected novels that were published in the last 50 years (1970–2020), with three novels per decade in each language, to capture a relatively more recent sample of motion event descriptions. We also included comparable numbers of novels written by male and female authors, as well as both award-winning and non-award-winning novels, to attain a more representative sample of written motion descriptions in literary texts in the three languages (see Appendix, for the list of novels in each language).

2.2 Data collection

Each novel was opened randomly ten times by a coder, with the goal to elicit a more representative sample of event descriptions across different parts of the book. At each opening the first observed instance of a physical motion event (e.g., the boy runs through the park) and a metaphorical motion event (e.g., the idea runs through the mind) was recorded, following the same procedure for extracting metaphorical motion events in earlier work (Özçalışkan 2003, 2005). A physical motion event was defined as the physical movement of a concrete entity (e.g., person) in relation to a physical location (e.g., girl jumps over the hurdle); a metaphorical motion was defined as the abstract motion of either a non-physical entity (e.g., time; years run by quickly) or the abstract motion of a physical entity in relation to an abstract location (e.g., mental states; she falls into despair). This resulted in 300 motion event descriptions (150 physical events, 150 metaphorical events) in each language and 900 motion descriptions across the three languages.

2.3 Data coding

All physical and metaphorical motion descriptions were segmented into sentence units. A sentence unit was defined as a segment of speech that contained at least one motion verb and its associated arguments and subordinate clauses, following earlier work (Özçalışkan et al. 2016a; e.g., The man slowly walks down the corridor; The feeling of relief runs through her body; The possibility enters his mind tiptoeing; The lady passes through the plaza in a rush). Some of the sampled instances from the novels included multiple sentence units, only one of which conveyed motion information (e.g., She was playing piano when the girl entered the house limping). For such instances, we only included the segment of the sampled instance that conveyed motion (e.g., when the girl entered the house limping) in our analysis given that our focus was on motion descriptions.

Each sentence unit was then coded for packaging of semantic information, following earlier work (Lewandowski and Özçalışkan 2018; Özçalışkan et al. 2016a). Sentence units were coded as either conflated, where path and manner components were both expressed within a single clause (e.g., the boy runs through the park; the idea runs through my mind) or separated, where a sentence unit contained only one motion component, namely path-only (e.g., the boy passes through the park; the idea passes through the mind), manner-only (e.g., the boy runs, the idea pops), or both components but in a main clause + subordinate clause construction (the boy passes the park running; the idea passes through the mind running; see Tables 1 and 2 for examples of each packaging type from our sample of written texts). Our decision to include main clause + subordinate clause constructions in the separated category was based on the extension of the motion event description over two clauses (instead of a single compact clause)—a coding decision that was consistent with all of the earlier work that focused on packaging of motion events (Akhavan et al. 2017; Fritz et al. 2019; Kita and Özyürek 2003; Lewandowski and Özçalışkan 2018; Özçalışkan et al. 2016a, 2016b, 2018; Tütüncü et al. 2023).

Next, we coded each sentence unit at the lexical level for both verbs and secondary motion expressions (i.e., linguistic elements outside the main verb root, such as particles, prepositions, prefixes, adverbs). For verbs, we coded the main verb of each sentence unit as expressing path (e.g., enter, ascend, cross) or manner (e.g., run, crawl, jump); we also coded each secondary motion expression as expressing path (e.g., into ecstasy, over the bridge) or manner (e.g., rapidly, at a slow pace), following earlier work (e.g., Özçalıskan 2009; Özçalıskan and Slobin 1999, 2003; see Tables 3 and 4 for examples of secondary motion expressions in the three languages).

Motion verbs that encoded both manner and path simultaneously (e.g., ‘sink’, ‘dive’) were extremely rare in our data (6 instances across 3 languages); we included these verbs under the manner category, following earlier work (e.g., Özçalıskan and Slobin 1999; Talmy 2000; see Hickmann et al. 2009; Hsiao 2009; Zlatev and Yangklang 2004 for further discussion of verbs that encode both components simultaneously).

We excluded 12 % of the sentence units expressing physical motion (M = 1.27, SD = 1.19, 57 instances) and 8 % of sentence units expressing metaphorical motion (M = 0.84, SD = 0.85, 38 instances) across the three languages from the lexicalization analysis as they conveyed motion with verbs that neither conveyed manner or path of motion (e.g., ‘move’, ‘go’). We also excluded a subset of these instances from the packaging analysis for both literal (6 %; M = 0.60, SD = 0.89, 27 instances) and metaphorical (5 %; M = 0.49, SD = 0.59, 22 instances) motion as they conveyed neither conflated nor separated packaging strategies (e.g., German: die Augen …, die sich bei Nacht bewegten ‘… the eyes … that moved at night …’ [Funke 2012, p. 158]; Polish: … autobus ruszył … ‘… the bus moved …’ [Stasiuk 1999, p. 4]; Spanish: … la fe mueve a las montañas … ‘… fatih moves mountains …’ [Puig 1976, p. 64]).

Reliability was assessed with three independent coders, each native speaker of one of the three languages of our study. The independent coders coded 20 % of the data, which included three randomly chosen writers in each language. Agreement between coders was 83 % for packaging of motion elements (separated, conflated) and 92 % for lexicalization of motion elements (manner vs. path verbs; secondary motion elements encoding manner vs. path).

2.4 Data analysis

Packaging: We first computed the total number of sentence units with conflated or separated packaging strategies for each novel and analyzed differences with two-way ANOVAs, with language as between (i.e., German, Polish, Spanish) and packaging type (conflated, separated) as within-subject factors—separately for physical motion and metaphorical motion descriptions.

Lexicalization: We next computed the total number of sentence units with motion verbs expressing manner or path information in each novel, separately for verb tokens and verb types. We analyzed differences with two-way ANOVAs, with language as between and verb type (manner, path) as within subject factors, separately for physical and metaphorical motion descriptions.

We also computed the total number of sentence units with one or more secondary motion expressions expressing manner or path in each novel. The incidence of secondary motion elements encoding manner was extremely rare in all three languages in the expression of both physical and metaphorical events; we therefore did not conduct any statistical analysis on this category of coding. We analyzed differences in the sentence units with secondary motion elements expressing path—which were frequently observed in our data—using one-way ANOVAs, with language as a between-subjects factor, separately for physical and metaphorical motion events. We used Bonferroni corrections for all follow-up pairwise comparisons.

3.1 Packaging of physical and metaphorical motion elements

We first examined packaging patterns in writers’ description of physical motion events, comparing German and Polish, both S-languages, to Spanish, a V-language. As can be seen in Figure 1A, our results showed no effect of language (F (2, 42) = 0.09, p = 0.92), but an effect of packaging type (F (1, 42) = 6.14, p = 0.018, η² _p  = 0.13), which interacted with language (F (2, 42) = 48.31, p < 0.001, η² _p  = 0.70). In line with our predictions, German and Polish writers used significantly more conflated descriptions than Spanish writers (Bonferroni, ps < 0.001). This pattern was reversed for Spanish writers, who relied more on separated descriptions compared to both German and Polish writers (Bonferroni, ps < 0.001). German and Polish novelists, on the other hand, did not differ from each other in their preference for separated or conflated responses (Bonferroni, ps = 1.0).

Figure 1:

Mean number of sentence units encoding physical (A) or metaphorical motion (B) with separated (light bars) or conflated (dark bars) packaging of motion elements in German, Polish, and Spanish; error bars represent standard error.

We next examined packaging patterns in the description of metaphorical motion events in the three languages. As Figure 1B illustrates, our results showed no effect of language (F (2, 42) = 0.84, p = 0.44), but an effect of packaging type (F (1, 42) = 20.16, p < 0.001, η² _p  = 0.32) that interacted with language (F (2, 42) = 69.71, p < 0.001, η² _p  = 0.77). Similar to physical motion events and in line with our predictions, German and Polish writers used significantly more conflated descriptions than Spanish writers (Bonferroni, ps < 0.001). Conversely, Spanish writers predominantly relied on the separated strategy, and did so more than both German and Polish writers (Bonferroni, ps < 0.001). German writers also differed from Polish writers, using fewer conflated (Bonferroni p < 0.001) but greater separated strategies (Bonferroni, p = 0.001) in their metaphorical motion descriptions.

Overall, the patterns of packaging in German, Polish, and Spanish written texts followed the inter-typological split in both physical and metaphorical motion event descriptions–with greater reliance on the conflated strategy by German and Polish novelists compared to Spanish novelists, and an almost exclusive preference for the separated strategy by Spanish novelists. This pattern was also accompanied by some intra-typological variation where German and Polish speakers differed in their relative production of each packaging strategy, but only in the context of metaphorical motion events.

3.2 Lexicalization of physical and metaphorical motion elements

3.2.1 Expression of manner and path in the verb

Beginning with physical motion descriptions with manner or path verbs, we found no effect of language (F (2, 42) = 0.75, p = 0.48), but an effect of verb type (F (1, 42) = 27.79, p < 0.001, η² _p  = 0.40), and more importantly, an interaction between language and verb type F (2, 42) = 62.15, p < 0.001, η² _p  = 0.75). As predicted, descriptions with manner verbs were higher in German and Polish novels compared to Spanish novels (Bonferroni, ps < 0.001). Spanish writers showed the reverse pattern and employed a greater number of descriptions with path verbs than both Polish and German (Bonferroni, ps < 0.001) writers. German and Polish novelists, however, did not differ in their token production of either descriptions with manner or path verbs (Bonferroni, ps ≥ 2.23; see Figure 2, panel A).

Turning next to metaphorical motion descriptions with manner or path verbs, we found largely similar patterns, with no effect of language (F (2, 42) = 0.75, p = 0.48) or verb type F (1, 42) = 1.89, p = 0.18) but a significant interaction between language and verb type F (2, 42) = 56.05, p < 0.001, η² _p  = 0.73). Following our prediction, the descriptions of metaphorical motion with manner verbs in German and Polish was higher compared to descriptions in Spanish (Bonferroni, ps < 0.001). The reverse was true for descriptions with path verbs: Spanish writers used greater number of sentence units with path verbs than their German and Polish counterparts (Bonferroni, ps < 0.001). German and Polish writers also differed from each other in their production of descriptions with both manner (Bonferroni, p = 0.001) and path (Bonferroni, p = 0.003) verbs, with German novelists producing fewer descriptions with manner verbs but more descriptions with path verbs than Polish novelists (Figure 2, panel B).

Figure 2:

Mean number of sentence units encoding physical motion (A) or metaphorical motion (B) with path (light bars) or manner (dark bars) verbs in German, Polish, and Spanish; error bars represent standard error.

The inter-typological differences were evident in the diversity of manner and path verbs (i.e., sentence units with different types of manner or path verbs) as well, with a significant interaction between language and verb type in writers’ description of both physical motion events (F (2, 42) = 48.96, p < 0.001, η² _p  = 0.70) and metaphorical motion events F (2, 42) = 42.63, p < 0.001, η² _p  = 0.67).^[3] As can be seen in Table 5, Spanish writers produced descriptions with fewer types of manner verbs than both German and Polish writers (Bonferroni, ps ≤ 0.001) in their description of both types of motion events. This pattern was reversed for path verbs, with more descriptions with different path verb types in novels in Spanish than the ones in either in German or Polish (Bonferroni, ps < 0.001). German and Polish authors were largely similar in their overall production of descriptions with the two verb types, with the only exception of German authors using a greater number of descriptions with different path verbs than their Polish counterparts in their metaphorical descriptions (Bonferroni, p = 0.009)—a difference that was not in line with our predictions; see Tables 6 and 7 for a complete list of the types of verbs, along with the relative frequencies, used by novelists in their description of physical and metaphorical motion events in each language.

Table 5:

Mean number of sentence-units with different types of manner and path verbs produced in the description of physical and metaphorical events in the three languages.

Physical motion	MANNER VERBS	PATH VERBS
	M (SD)	M (SD)
German	5.33 (1.35)	2.07 (0.70)
Polish	4.73 (1.75)	1.53 (1.06)
Spanish	2.20 (0.86)	5.27 (1.44)
Metaphorical motion
German	4.40 (1.06)	3.0 (1.20)
Polish	4.93 (1.79)	1.73 (0.96)
Spanish	1.87 (1.46)	5.53 (1.13)

Table 6:

Manner and path verbs used in physical motion descriptions by novelists in the three languages. Numbers in parentheses indicate frequency of use for each verb.

MANNER VERBS	PATH VERBS
German	German
begleiten ‘accompany’ (1), bohren ‘drill’ (1), fahren ‘drive, go by car’ (12), fliehen ‘escape’ (1), führen ‘lead’ (3), gehen ‘walk/go’ (20), gießen ‘pour’ (1), laufen ‘run’ (2), pressen ‘press’ (2), rasen ‘race’ (1), reißen ‘tear’ (2), reiten ‘ride on horse’ (1), rennen ‘run/race’ (3), rieseln ‘trickle’ (1), rutschen ‘slide’ (1), saugen ‘suck’ (1), schieben ‘push’ (2), schubsen ‘push/nudge’ (1), schütteln ‘shake’ (4), schwanken ‘sway’ (1), Spaziergänge machen ‘go for walks’ (1), springen ‘jump’ (3), spülen ‘rinse’ (1), stapfen ‘stomp’ (1), stoßen ‘bump’ (1), streifen ‘streak’ (1), strömen ‘flow’ (1), tragen ‘carry’ (4), treten ‘step’ (4), wandern ‘hike, wander’ (2), wanken ‘stagger’ (1), werfen ‘throw’ (6), ziehen ‘pull’ (4)	bringen ‘bring’ (7), (sich) erheben ‘raise/lift’ (2), erreichen ‘reach’ (1), fallen ‘fall’ (3), folgen ‘follow’ (1), holen ‘fetch’ (2), kommen ‘come’ (13), steigen ‘go up’ (1), überqueren ‘cross’ (1), verlassen ‘leave’ (5), zurückkehren ‘come back’ (3)
Polish	Polish
biec ‘run’ (4), chodzić/iść ‘walk’ (44), chwiać się ‘wobble’ (1), ciągnąć ‘pull’ (2), cisnąć się ‘squeeze through’ (2), fruwać ‘flutter’ (1), jechać ‘drive, go by car’ (8), kołysać ‘sway’ (1), kroczyć ‘tread’ (2), lać ‘pour’ (1), latać ‘fly’ (1), lądować ‘land’ (1), lgnąć ‘cling’ (1), nosić ‘carry’ (5), pędzić ‘race’ (2), pluć ‘spit’ (1), pływać ‘swimm’ (1), prowadzić ‘lead’ (1), przedzierać się ‘wade’ (1), przemycać ‘smuggle’ (1), rwać ‘tear’ (3), rzucać ‘throw’ (4), skakać ‘jump’ (4), sunąć ‘glide’ (2), ślizgać się ‘slide’ (2), trzęść się ‘shake’ (2), uciec ‘escape’ (1), wozić ‘drive, carry by car’ (2), wlec ‘drag’ (1), wywijać ‘wave’ (1), zanurzyć ‘dip’ (3)	cofać się ‘step back’ (1), od/prze-wracać (się) ‘turn around’ (2), krążyć ‘circle’ (3), mijać ‘pass’ (1), opuszczać ‘leave’ (2), padać ‘fall’ (5), wracać ‘come back’ (8), wyjąć ‘take out’ (4)
Spanish	Spanish
andar ‘walk’ (4), arrancar ‘tear’ (1), arrojar ‘throw’ (1), balancearse ‘swing’ (1), callejear ‘loiter’ (1), caminar ‘walk’ (1), conducir ‘drive/lead’ (4), correr ‘run’ (1), dar un paso ‘take a step’ (1), deslizarse ‘slide’ (1), echar ‘throw’ (1), escapar ‘escape’ (1), escurrir ‘drain’ (1), flotar ‘float’ (1), huir ‘escape’ (1), hundir ‘sink’ (1), manejar ‘drive’ (1), perseguir ‘chase’ (1), resbalar ‘slide’ (1), saltar ‘jump’ (1), soltar ‘drop’ (1), temblar ‘shake’ (1), tirar ‘throw’ (2), tropezar ‘stumble’ (1), volar ‘fly’ (1)	acercarse ‘approach’ (2), acudir ‘come to a place’ (5), alejar(se) ‘move away’ (4), apartar(se) ‘move away’ (1), apearse ‘get off’ (1), aproximarse ‘approach’ (1), avanzar ‘move along’ (2), bajar ‘descend’ (1), caer ‘fall’ (3), cruzar ‘cross’ (4), dar vueltas ‘circle’ (1), desprenderse ‘come off’ (1), entrar ‘enter’ (7), irrumpir ‘burst in’ (1), llegar ‘arrive’ (6), llevar ‘take (away from a place)’ (2), meter ‘put in’ (4), pasar ‘go past’ (4), pirarse ‘col. leave’ (1), regresar ‘return (5)’, retirar ‘remove/take away’ (1), sacar ‘take out’ (1), salir ‘exit’ (14), seguir ‘follow’ (2), traer ‘bring’ (5), venir ‘come’ (7), volver ‘return’ (5)

Table 7:

Manner and path verbs used in metaphorical motion descriptions in German, Polish and Spanish. Numbers in parentheses indicate frequency of use for each verb.

MANNER VERBS	PATH VERBS
German	German
begleiten ‘accompany’ (2), dringen ‘force one’s way through’ (1), fahren ‘drive, go by car’ (3), fliegen ‘fly’ (2), fließen ‘flow’ (1), führen ‘lead’ (7), gehen ‘walk/go’ (18), hasten ‘hurry’ (1), huschen ‘scurry’ (1), laufen ‘run’ (3), münden ‘flow into’ (1), ringen ‘wrestle’ (1), schieben ‘push’ (2), schwimmen ‘swimm’ (1), sinken ‘sink’ (3), springen ‘jump’ (2), stolpern ‘stumble’ (1), strömen ‘flow’ (1), stürzen ‘hurl’ (1), tanzen ‘dance’ (1), tauchen ‘dive’ (3), tragen ‘carry’ (7), treiben ‘drive (animals)’ (1), treten ‘step’ (3), wandern ‘hike, wander’ (1), wehen ‘blow’ (1), werfen ‘throw’ (4), ziehen ‘pull’ (2)	bringen ‘bring’ (9), sich entfernen ‘move away’ (2), (sich) erheben ‘raise/lift’ (2), erreichen ‘reach’ (1), erscheinen ‘appear’ (1), fallen ‘fall’ (12), folgen ‘follow’ (3), geraten ‘get/come somewhere’ (3), holen ‘fetch’ (1), kommen ‘come’ (22), umgeben ‘surround’ (1), verschwinden ‘disappear’ (1)
Polish	Polish
biec ‘run’ (6), chodzić/iść ‘walk’ (42), ciągnąć ‘pull’ (6), cisnąć się ‘squeeze through’ (1), dźwigać ‘carry’ (1), gonić ‘chase’ (1), kołysać ‘sway’ (1), kroczyć ‘tread’ (1), kuleć ‘limp’ (1), latać ‘fly’ (2), miotać ‘hurl’ (1), nosić ‘carry’ (4), pędzić ‘race’ (1), pełzać ‘creep’ (2), pływać ‘swimm’ (5), prowadzić ‘lead’ (6), pryskać ‘splash’ (1), rwać ‘tear’ (3), rzucać ‘throw’ (6), skakać ‘jump’ (4), stąpać ‘tread’ (1), sunąć ‘glide’ (1), sypać ‘spill’ (1), topić ‘drown’ (1), uciec ‘escape’ (2), wędrować ‘hike, wander’ (1), włóczyć (się) ‘roam’ (1), wozić ‘drive, carry by car’ (1), zwalić ‘knock off’ (1)	dotrzeć ‘reach’ (1), krzyżyować ‘cross’ (1), obracać ‘turn around’ (1), mijać ‘pass’ (5), opuszczać ‘leave’ (2), padać ‘fall’ (8), podwyższać ‘raise’ (1), wracać ‘come back’ (7), wzbić się ‘rise up’ (1)
Spanish	Spanish
acompañar ‘accompany’ (2), arrancar ‘tear’ (1), correr ‘run’, derramar ‘spill’ (1), echar ‘throw’ (3), empujar ‘push’ (1), escapar ‘escape’ (3), esparcir ‘scatter’ (1), flotar ‘float’ (1), huir ‘escape’ (1), hundir ‘sink’ (3), lanzar ‘launch’ (4), rondar ‘patrol’ (1), saltar ‘jump’ (3), soltar ‘drop’ (1), tambalearse ‘sway’ (1), tirar ‘throw’ (1), vagar ‘roam’ (2)	acceder ‘access’ (1), abandonar ‘leave’ (1), alcanzar ‘reach’ (2), alzar ‘raise/lift (2), alejar(se) ‘move away’ (1), bajar ‘descend’ (1), caer ‘fall’ (8), cruzar ‘cross’ (3), descender ‘descend’ (1), entrar ‘enter’ (8), llegar ‘arrive’ (7), llevar ‘take (away from a place)’ (8), meter ‘put in’ (3), pasar ‘go past’ (24), quitar ‘take out’ (1), sacar ‘take out’ (6), salir ‘exit’ (4), seguir ‘follow’, subir ‘ascend’ (3), traer ‘bring’ (1), venir ‘come’ (3), volver ‘return’ (5)

3.2.2 Expression of manner and path in secondary motion elements outside the verb

In physical motion descriptions, expression of manner outside the verb was extremely rare in all three languages, but all three languages frequently expressed path information outside the verb. In line with our predictions, writers showed an effect of language (F (2, 42) = 15.71, p < 0.001, η² = 0.43). In contrast to our predictions, German and Polish writers were comparable in their production of descriptions with secondary motion elements conveying path of motion (p = 1.0). However, in line with our predictions, both differed from Spanish writers, who used fewer sentence units with secondary path expressions (see Figure 3A; Bonferroni, ps < 0.001).

Figure 3:

Mean number of sentence units encoding physical motion (A) or metaphorical motion (B) with secondary motion expressions conveying path (light bars) or manner (dark bars) information in German, Polish, and Spanish; error bars represent standard error.

The pattern was similar in metaphorical motion descriptions: writers in each language used secondary manner expressions extremely rarely. In contrast, the incidence of secondary path expressions was much more frequent, but also showed an effect of language (F (2, 42) = 47.93, p < 0.001, η² = 0.70). As can be seen in Figure 3B, German and Polish speakers produced comparable number of descriptions with secondary path expressions (Bonferroni, p = 0.49), but both differed from Spanish speakers, who used fewer of such descriptions (Bonferroni, ps < 0.001).

In summary, the lexicalization patterns in German, Polish, and Spanish novels showed strong inter-typological differences between S- and V-languages for both physical and metaphorical motion events. German and Polish writers (both S-languages) produced descriptions with greater amount (tokens) and variety (types) of manner verbs than Spanish (V-language) writers who, in turn, produced descriptions with higher amount and variety of path verbs. These differences were also evident in secondary motion expressions conveying path of motion: German and Polish novelists frequently used secondary expressions conveying path of motion and at rates greater than Spanish novelists. Our results also suggested some evidence for intra-typological differences in the lexicalization of motion events between German and Polish novelists in their use of manner and path verbs—but only in the context of metaphorical motion events.

4.1 Inter-typological variability in the expression of physical and metaphorical motion events

In describing physical and metaphorical motion, authors followed the expected inter-typological differences in their packaging and lexicalization of motion. German and Polish writers used more conflated descriptions, with greater number and variety of manner verbs than Spanish writers. This pattern was reversed for Spanish writers, who showed greater preference for the separated pattern, with greater amount and variety of path verbs compared to their German and Polish counterparts. Our study extends earlier work that showed similar patterns of inter-typological variability for both the packaging (e.g., Kita and Özyürek 2003; Özçalışkan et al. 2016a, 2016b) and the lexicalization (e.g., Cardini 2010; Hickmann et al. 2009; Özçalışkan 2009; Tusun and Hendriks 2019, 2022) of physical motion events elicited orally by experimental probes. Our findings also align with earlier work that showed the same patterns of crosslinguistic variability in the lexicalization of metaphorical motion events conveyed in written texts (Caballero and Ibarretxe-Antuñano 2015; Cifuentes-Férez 2014; Özçalışkan 2003, 2004, 2005) and oral elicitations (Katis and Selimis 2005; Özçalışkan 2005, 2007; Wilson 2005) in other S- (i.e., English) and V-languages (i.e., Greek, Spanish, Turkish). Moreover, the close resemblance in patterns of inter-typological variability in the two event types was evident not only across different target domains (as shown in our analysis) but also in each target domain, with greater use of conflated packaging strategy and manner verbs in German and Polish (S-languages) as compared to Spanish (V-language) in the expression of both physical and metaphorical motion descriptions. This finding contrasts with some of the earlier work that showed lower expression of manner in the metaphorical descriptions of time (Feist and Duffy 2020) and perceptual events (i.e., fictive motion; Rojo and Valenzuela 2003) in S-languages (i.e., English), but is consistent with several others that showed similar patterns of inter-typological variability in the metaphorical motion of different target domains across different languages (Katis and Selimis 2005; Özçalışkan 2005).

Importantly, the inter-typological patterns were also observable at the level of secondary motion expressions for both physical and metaphorical motion events. In fact, our study joined others (e.g., Cadierno and Ruiz 2006; Hickmann et al. 2009; Özçalışkan 2004, 2005; Özçalışkan and Slobin 2003), showing very infrequent incidence of secondary manner expressions, but frequent occurrence of secondary path expressions that followed inter-typological lines for both event types. These findings thus suggest that patterns of inter-typological variability are robust and become observable across different S- versus V-languages as well as different motion event types (i.e., metaphorical vs. literal).

Why do patterns of inter-typological variability remain similar in both physical and metaphorical event types? One explanation could be the way a metaphorical event is structured: a metaphorical event involves an abstract target domain (e.g., ideas) which is structured by a source domain that is closely tied to physical experience (i.e., motion; Lakoff and Johnson 1980, 1999). As such, the linguistic patterns that are evident in the expression of the physical experience that constitutes the source domain are largely reflected in the expression of the metaphorical concept. Another explanation could be tied to the ‘simulation account’ of metaphorical event processing (e.g., Gibbs 1994; Matlock 2004), which suggests that speakers (or writers, as in our study) mentally process metaphorical motion events by simulating physical motion events. Earlier studies that used a range of techniques from reaction time (Matlock 2004, 2006) and eye-tracking (e.g., Richardson and Malock 2007) to neural imaging (e.g., Saygın et al. 2009) showed strong evidence for the close alignment between the processing of physical and metaphorical motion events. As such, the close alignment in the inter-typological patterns of linguistic expression in physical and metaphorical events that we observed in our study might also stem from the similar ways speakers process both event types in a given language.

One intriguing question is why the inter-typological patterns of variability are pervasive across multiple contexts of language use, from describing everyday physical events (e.g., a boy crawling over a carpet) to expressing metaphorical concepts (e.g., an idea creeping into one’s mind) across a variety of languages. One potential explanation for the strength of the inter-typological differences across tasks of varying complexity could be based on developmental research that has examined changes in children’s expression of motion events. Earlier developmental work that focused on events involving physical motion (e.g., Allen et al. 2007; Choi and Bowerman 1991; Filipović 2007; Hickmann et al. 2009; Özçalışkan and Slobin 1999; Özçalışkan et al. 2023) or metaphorical motion (Özçalışkan 2007) showed that patterns of inter-typological variability emerge at an early age (∼ages 3–5) and remain largely unchanged over developmental time, regardless of changes in the complexity of the child’s developing language system. These findings thus suggests that inter-typological constraints on the linguistic construal of motion operate with equal strength, regardless of the degree of complexity (i.e., from elicited description of everyday events to creative writing about events) with which a speaker expresses that event.

Another intriguing question is to what extent expression of metaphorical motion might be distinguishable from expression of physical motion, particularly given the relative lack of recognition of metaphor in everyday encounters. The pervasiveness of metaphors in general and metaphorical motion events in particular in everyday discourse makes it unlikely for metaphorical motion to be recognized as distinct from physical motion events, particularly in adult first language production contexts. In fact, there is evidence that suggests that adult speakers process metaphorical motion by mentally simulating representations of physical motion (Gibbs 1994; Matlock 2004). The strong differences in the expression of metaphorical motion we observed in our data thus raise the possibility that adults might also think about the same abstract concepts in different ways depending on what language they speak: perhaps ideas are imagined more as jumping, wandering or running in German and Polish, while they are seen as only passing in Spanish without any manner. We know from previous work that crosslinguistic differences in the expression of physical motion influence the way speakers think about such events, particularly when the tasks involve verbalization of the event (Emerson et al. 2021; Gennari et al. 2002; Hohenstein 2005; Özçalışkan et al. 2016a). Future work that examines the effects of crosslinguistic variability on speakers’ cognition about metaphorical events can determine whether language-specific effects on cognition extend to metaphorical motion events.

4.2 Intra-typological variability in the expression of physical and metaphorical motion events

In describing physical and metaphorical motion, authors also showed some intra-typological differences in their lexicalization of motion, but only in the context of metaphorical motion events. Specifically, Polish writers used a greater number of conflated descriptions with manner verbs as compared to German writers, who, in turn, produced more separated descriptions with path verbs. Contrary to our predictions, however, there was no difference in the diversity of manner verbs (verb types) between German and Polish in either event type.

What explains the higher incidence of conflated descriptions and manner verbs in Polish? As shown in earlier work (Lewandowski 2010, 2014), Slavic languages have a smaller inventory of high frequency path verbs, particularly when compared to Germanic languages—a pattern that was evident in our data as well. In expressing path in the verb, German speakers relied predominantly (52 % of path verbs) on two deictic verbs: kommen ‘come’, bringen ‘bring’—neither of which have direct equivalents in Polish. Polish writers, in contrast, relied on path prefixes to convey similar deictic trajectories, frequently combining them with manner verbs (e.g., przy-jść ‘toward-walk’, przy-jechać ‘toward-drive/go by car’, przy-nieść ‘toward-carry [walking]’), thus resulting in increased rates of conflated strategies and manner verb tokens in Polish compared to German. These findings, in fact, extend previous work on oral descriptions of physical motion, which showed similar patterns of intra-typological variability in the relative use of manner and path verbs in Germanic (e.g., English, German) versus Slavic languages (e.g., Polish, Russian; Lewandowski 2021; Slobin 2006).

But why did we not observe the previously reported intra-typological differences in manner verb types that showed an advantage for German writers? One likely reason could be the elicitation method used in earlier studies. Previous evidence showing greater use of manner verb types in German (as compared to Polish) was largely based on descriptions of visual scenes (picture books, video animations) with salient manners (i.e., flipping over a beam, tumbling into a trashcan; e.g., Bamberg 1994; Lewandowski and Özçalışkan 2021; Ochsenbauer and Hickmann 2010). The elicited descriptions of such events might have cued speakers to pay greater attention to manner and consequently include a greater variety of manner distinctions in their expression of motion. This effect might have been particularly pronounced given the structure of the two languages: the morphologically free satellites in German are compatible with a wide range of manner verbs, whereas the morphologically bound satellites in Polish only admit a limited set of basic manner verbs (Lewandowski 2021; Lewandowski and Mateu 2020), thus resulting in greater variety of manner verbs in German than in Polish in the earlier studies that relied on manner-salient stimuli. Our study, by contrast, relied on more naturalistic production of event descriptions by writers, without the use of manner-salient probes. Accordingly, the differences in task demands might have resulted in lack of reliable intra-typological differences in the relative production of manner verb types in the two languages.

But why did only metaphorical but not physical motion descriptions display intra-typological variability? In our study, German writers tended to use more path verbs in metaphorical than in physical motion descriptions—a trend that was not observable in Polish, largely because of its limited inventory of path verbs. The slight increase in path verbs relative to manner verbs in metaphorical descriptions in German might be the underlying factor in detecting intra-typological differences in metaphorical but not physical motion descriptions. Our findings thus suggest that the level of intra-typological variability might depend on the abstractness of the event (physical vs. metaphorical)—with greater ‘abstractness’ associated with greater intra-typological variability.

4.3 Conclusions

Prior crosslinguistic work on motion events largely focused on inter-typological variability in oral descriptions of physical motion produced by adult native speakers (e.g., Kita and Özyürek 2003; Lewandowski and Özçalışkan 2021; Naigles et al. 1998; Özçalışkan et al. 2016a, 2018; Tusun and Hendriks 2019, 2022) and showed ample evidence for distinct patterns of motion event expression in S- versus V-languages—both in terms of broader packaging strategies (separated vs. conflated) and more specific patterns in the lexicalization of path and manner inside and outside the verb. Our study extends these findings to a relatively understudied production context—the literary genre—and to another type of event, namely metaphorical motion, approaching it from both inter-typological and intra-typological perspectives.

Our results showed largely similar inter-typological patterns in the way speakers packaged and lexicalized physical and metaphorical motion events. At the same time, the close inter-typological alignment was accompanied by some intra-typological variation, but only in the metaphorical domain. Overall, our results suggest that inter-typological variation remains stable across event types, but intra-typological variation is less robust, as it plays out with differential strength, depending on type of task (e.g., experimental vs. naturalistic data) and type of event (e.g., physical vs. metaphorical).