1. Introduction
Local prepositions are used to express spatial relations (e.g., Das Buch liegt auf dem Tisch. “The book is lying on the table.”). Other prepositions are temporal, modal, or causal prepositions. German children first acquire local prepositions, followed by temporal ones (e.g., Grimm, Reference Grimm1975; Harr & Scholtz, Reference Harr and Scholtz2014; Tomasello, Reference Tomasello1987). To produce or understand sentences correctly that involve prepositions, children and adults also need to have knowledge about case marking and its appropriate use. The German language is characterized by four different cases: nominative, accusative, dative, and genitive. This paper deals with case marking with prepositions. It will not discuss the intricacies of case assignment, but note here that there are other case assigners in German, e.g., verbs. In German, there are prepositions that assign a unique case, so-called one-case prepositions, and there are prepositions that can assign two cases, so-called two-case prepositions. In total, there are nine local two-case prepositions – an “at,” auf “onto, on,” hinter “behind,” in “into, in,” neben “next to,” über “over, above,” unter “under,” vor “in front of,” and zwischen “between” – which can assign accusative and dative case depending on the intended meaning. The different syntactic structure results in a difference in meaning. If a two-case preposition assigns accusative case to the following determiner phrase (DP), a directional meaning results (see (2)), and if it assigns dative case, a static interpretation results (see (3)) (Tracy, Reference Tracy1986). In (1), zu “towards” is a one-case preposition assigning dative case. In “into, in” is a two-case preposition that can assign accusative (see (2)) and dative case (see (3)). The interpretation in (2) is directional as a change of place is indicated (with the chick going from being outside the box to being inside it). In (3), the interpretation is static, indicating no change of place (with the chick already being in the box and the movement taking place in the box).
As the German case marking system consists of many different forms with no one-to-one form–meaning mapping, the adult-like use of case marking is challenging for children learning German (e.g., Meisel, Reference Meisel1986; Ulrich et al., Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016). For children, marking of case assigned by prepositions is considered harder than marking of case assigned by verbs (e.g., Mills, Reference Mills and Slobin1985). Besides the meaning of the preposition, children need to learn which case is assigned by each individual preposition. For one-case prepositions, there is one case to store with the corresponding preposition. Two-case prepositions are difficult due to their conceptual complexity as they allow the use of accusative and dative, thus triggering different sentence meanings (Tracy, Reference Tracy1986), as illustrated in (2) and (3).
Being able to understand and use the correct case marking to trigger the intended meaning is important before school entry, as it is essential to the correct interpretation of a sentence (Ulrich et al., Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016). Comprehending and producing the correct case and thus processing morpho-syntactic rules not only are necessary for sentence production and comprehension but might be seen as a predictor for understanding written texts (e.g., Carlisle, Reference Carlisle and Feldman1995; Deacon & Kirby, Reference Deacon and Kirby2004; Jongejan et al., Reference Jongejan, Verhoeven and Siegel2007; Kuo & Anderson, Reference Kuo and Anderson2006), as well as for understanding narratives and developing narrative skills (see Mokhtari & Thompson, Reference Mokhtari and Thompson2006; Nation & Snowling, Reference Nation and Snowling2000), which have been described as aspects of morphological and syntactical awareness/grammar.
1.1. Theoretical review of case and preposition acquisition in monolingual children
There are various approaches explaining why one case (e.g., accusative) is learned before another (e.g., dative). In comparison, only few explanations deal with the acquisition of one-case and two-case prepositions. In this section, we will discuss theories on the acquisition of accusative and dative case and on one-case and two-case prepositions.
There is consensus on the order of acquisition of case marking with verbs: first nominative case, then accusative case, and finally dative case (see Clahsen, Reference Clahsen1984; Tracy, Reference Tracy1986).
Why the different German cases are acquired in a certain order has been explained by different theories. The different case marking on the determiner in relation to the noun’s gender is illustrated in Table 1.
Table 1. Overview of the German nominative, accusative, and dative case marking on the definite article
According to the theory of morphological marking (Fliedl, Reference Fliedl1999; Smits et al., Reference Smits, Mortelmans and Willems2020), morphologically less marked forms are acquired before strongly marked forms. In comparison to nominative case, Table 1 shows that for accusative and nominative case, feminine nouns share the same determiner (die), as do neuter nouns (das). This comparison shows that case marking on the determiner is visible in masculine nouns as there are distinct determiners for nominative (der) and accusative (den) case. Hence, nominative case is considered unmarked, whereas accusative case is less marked. As for dative case, in comparison to nominative, masculine, feminine, and neuter nouns do not share the same determiner (masculine nouns der in contrast to dem, feminine nouns die in contrast to der, and neuter nouns das in contrast to dem). Therefore, dative case is considered strongly marked. Morphological marking provides an explanation for why dative case is acquired later than accusative case.
Besides morphological marking, input frequency can explain the acquisition of case. Wegener (Reference Wegener1995) found dative case to occur more frequently with prepositions than accusative case. If input can explain the acquisition of accusative and dative case assigned by prepositions, dative case should be learned before accusative case with prepositions due to its more frequent occurrence.
Baten and Willems (Reference Baten and Willems2012) explain the acquisition of one-case and two-case prepositions by the processability theory (PT). PT is a theory for second language acquisition. We will only discuss their argument of feature unification as this is part of theories on first language acquisition (most notably lexical -functional grammar; see, e.g., Börjars et al., Reference Börjars, Nordlinger and Sadler2019) as well. Features that are unified and transferred within or across a phrase are grammatical features like case, number, or gender. Within PT, Baten and Willems (Reference Baten and Willems2012) describe three phases of feature unification:
Phase 1: “No feature unification,” which means that children produce grammatically incorrect phrases as they do not yet possess the ability to unify grammatical information, e.g., *das Kiste, where the article is neuter and the noun is feminine, being therefore ungrammatical.
Phase 2: “Phrase-internal feature unification.” As Figure 1 shows, the grammatical information “case” is transferred within the prepositional phrase (PP). The preposition zu “towards” holds the case information dative case, which is transferred to the DP within the PP, so that the determiner der “the.DAT” is marked with dative case. It is unnecessary to go beyond the PP (framed in red). The PP includes all the required information. The grammatical information is thus unified phrase-internally. One-case prepositions are assigned to Phase 2.
Figure 1. Phrase-internal feature unification (adapted from Baten & Willems, Reference Baten and Willems2012). The red frame represents the internal transfer of the feature case (= dative) within the prepositional phrase (PP). Note: S = sentence; DP = determiner phrase; D = determiner; NP = noun phrase; IP = inflectional phrase; I = inflection; VP = verbal phrase; V = verb; PP = prepositional phrase; P = preposition; ACC = accusative; DAT = dative.
Phase 3: “Phrase-external feature unification,” which means features are transferred across phrases. Two-case prepositions are assigned to the third phase. Figure 2 depicts the transfer across the PP to the verbal phrase (VP). The preposition auf “onto, on” stores the grammatical information accusative and dative case, depending on the intended meaning. In this case, the features must be transferred across the PP. The combination of the information of the VP and the PP provides the necessary information. The verb sitzen “to sit” together with the preposition auf “onto, on” holds the information dative case. Consequently, the features are transferred phrase-externally, from the PP (framed in red) to the VP (framed in green) and vice versa.
Figure 2. Phrase-external feature unification. The red frame represents the transfer of the feature case (= dative and accusative) within the prepositional phrase (PP), whereas the green frame represents the transfer of the feature case (= dative) within the verbal phrase. The overlap of both frames shows that the feature case is transferred across phrases, leading to dative case assignment. Note: S = sentence; DP = determiner phrase; D = determiner; NP = noun phrase; IP = inflectional phrase; I = inflection; VP = verbal phrase; V = verb; PP = prepositional phrase; P = preposition; ACC = accusative; DAT = dative.
This approach of elucidating the distinction between one-case and two-case prepositions by using the syntactic structure can be embedded within Garrett’s (Reference Garrett and Butterworth1980) sentence processing account for production. Garrett delineates three distinct levels: (i) positional level, (ii) functional level, and (iii) message level. The positional level corresponds to the sentence structure depicted in Figures 1 and 2. The functional level is associated with grammatical categories and their semantic representation, while the message level signifies the speaker’s intention and the meaning of the message they wish to convey. In order to provide a full explanation of the uniqueness of two-case prepositions, it is necessary to add conceptual information to the structural planning, which is shown in Figure 3 by using the examples given in (2) and (3): Das Küken läuft in die Box. “The chick is walking into the box.” on the left and Das Küken läuft in der Box. “The chick is walking in the box.” on the right.
Figure 3. Sentence structure for sentences involving a two-case preposition together with the corresponding conceptual level. Note: S = sentence; DP = determiner phrase; D = determiner; NP = noun phrase; IP = inflectional phrase; I = inflection; VP = verbal phrase; V = verb; PP = prepositional phrase; P = preposition; ACC = accusative; DAT = dative.
As illustrated in Figure 3, the combination of the verb laufen “to walk” and the preposition in “into, in” does not provide any indication of which case should be used. Accusative and dative case are appropriate choices, and both result in grammatically correct sentences. The sentence structure alone does not reveal the correct case; it is imperative that children and adults master the conceptual information they wish to convey in order to correctly use the appropriate case marking.
In this section, we reviewed how the acquisition of accusative and dative case is contingent on the reliance on morphological marking or input frequency. The process of feature unification in conjunction with the conceptual level was proposed as a potential explanation for the intricacy of learning two-case prepositions.
1.2. Acquisition of case marking with prepositions in monolingual children
In this section, an overview of the existing literature on accusative and dative case marking with prepositions is outlined, taking into account production studies and studies on comprehension.
Table 2 provides an overview of the development of case marking with verbs and prepositions in monolingual German-speaking children. In production, case marking with verbs by monolingual German-speaking children was first described by Clahsen (Reference Clahsen1984). He reported three phases children go through while acquiring the case marking system. In Phase I, no case marking is observable. Children produce utterances without a determiner (e.g., gleich wauwau suche, Clahsen, Reference Clahsen1984, p. 7). In Phase II, children use case-neutral marking (e.g., Ich sehe ein Kater. “I see a cat.” instead of Ich sehe einen Kater. “I see a.ACC cat.”). In Phase III, case marking skills develop, first accusative case (Phase IIIa) and then dative case (Phase IIIb). Tracy’s (Reference Tracy1986) results support his findings. Case marking with verbs and with prepositions was studied in a monolingual child by Schmitz (Reference Schmitz, Hole, Meinunger and Abraham2006). The child already marked dative case assigned by prepositions correctly at age 4;6. Observable errors were article omission, no case marking, and overgeneralization of accusative case. The studies mentioned above analysed spontaneous speech of children between 1;0 and 4;10.
Table 2. Literature review of case marking with verbs and prepositions in children learning German as their first language
Note: DP = determiner phrase; PP = prepositional phrase; 2cP = two-case prepositions; 1cP = one-case prepositions; NOM = nominative; DAT = dative; ACC = accusative; > = more accurate.
Case marking with two-case prepositions was analysed by Tracy (Reference Tracy1986). The children in her sample had particular difficulties with two-case prepositions and overgeneralized accusative case.
To sum up, the German case marking system has been described as being acquired by age four. More recent research testing case marking in controlled experimental settings has revealed that especially the acquisition of dative case marking takes more time than expected (e.g., Ulrich et al., Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016).
Turgay (Reference Turgay2011) analysed accusative and dative case marking with one-case and two-case prepositions. She conducted four experiments to elicit case marking with prepositions (see Table 2). Monolingual children produced 83.9% of all PPs correctly, with the older children (age range: 7;0–10;11 years) performing more accurately than the younger ones (3;0–4;11 years). In 87.4% of all cases, children used the correct case with prepositions. Errors were no case marking (5%) and using the wrong case (7.4%), mostly marking accusative case instead of dative case (53.6%). The children performed less accurately in the production tasks than in the comprehension tasks. With one-case prepositions that assign accusative case, no difficulties were found. In 9.4% of the cases, they made mistakes with one-case prepositions that assign dative case. Concerning two-case prepositions, they obtained lower scores in dative case (error rate of 11.6% versus error rate of 3.1% for accusative case). Turgay concluded that there are no differences in marking case assigned by one-case and two-case prepositions as dative case was more difficult for children than accusative case with both types of prepositions.
Ulrich et al. (Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016) and Ulrich (Reference Ulrich2017) elicited accusative and dative case marking with verbs and prepositions in 968 monolingual children from 4;0 to 8;11 years. To test case marking with prepositions, the preposition hinter “behind” was used. Of the children tested, 218 were preschoolers. Applying the criterion that marking a case correctly in 90% of all obligatory contexts defines mastery of that case, the authors state that 28.9% of the children have fully acquired dative case and 40.8% of the five-year-olds have fully acquired accusative case. They marked accusative case correctly in 83% of occurrences and dative case in around 62.5%. With regard to the entire sample of 968 children, accurate production of the accusative case assigned by a verb occurred in 91.7% of all instances, which is higher than the 75.8% accuracy observed in the marking of the accusative case following a preposition. The dative case assigned by verbs was marked correctly in 68.3% of all instances, while the marking of the dative case assigned by a preposition achieved an accuracy rate of 83.4%. The difficulties in marking dative case were still observable in nine-year-old children. As posited by Ulrich et al. and Ulrich, children in all age groups had fully acquired dative case but not accusative case. The researchers thus conclude that the order of acquisition of nominative case first, followed by accusative case and dative case, does not hold true for all children.
Studies on case marking with two-case prepositions describe accusative as being acquired before dative (Tracy, Reference Tracy1986; Turgay, Reference Turgay2011). Regarding differences in case marking with one-case and two-case prepositions, Turgay’s (Reference Turgay2011) results show that children are equally accurate with one-case and two-case prepositions.
While understanding of thematic role assignment and cues in canonical and non-canonical sentences has been investigated extensively (Biran & Ruigendijk, Reference Biran and Ruigendijk2015; Gamper, Reference Gamper2016, Reference Gamper2019; Lindner, Reference Lindner2003; Schaner-Wolles, Reference Schaner-Wolles1989; Schipke et al., Reference Schipke, Knoll, Friederici and Oberecker2012, Reference Schipke, Stegenwallner-Schütz and Adani2024; Thelen, Reference Thelen2019; Watermeyer & Kauschke, Reference Watermeyer and Kauschke2013), less work has focused on children’s comprehension of case marking with prepositions. Two studies have analysed the understanding of case marking with two-case prepositions in monolingual German-speaking children. Paprotté (Reference Paprotté1979) used an act-out task to evaluate the comprehension of the two-case prepositions in “in,” auf “onto, on,” and unter “under.” The children were asked to put something into, onto, or under something. This means that one condition, the directional interpretation, was tested, not both potential interpretations. Older children performed more accurately than younger children. Children between 4;5 and 6;1 achieved accuracy rates of up to 100% in understanding in “into” and auf “onto,” and up to 91% for unter “under.” Turgay (Reference Turgay2011) administered two act-out tasks. The first task was intended to show whether children understand the meaning of the four prepositions neben “next to,” vor “in front of,” unter “under,” and hinter “behind.” They were asked to put a crocodile next to, in front of, under, and behind a box. With the second task, the understanding of accusative and dative case marking with the prepositions in “into” (accusative) or “in” (dative) and auf “onto” (accusative) or “on” (dative) was assessed. The children were asked to make the crocodile jump onto/on or into/in a box. Monolingual children incorrectly acted out the demanded action in the second task in 1.6% instances by making the crocodile jump onto or into a box instead of on or in one.
Summing up, research on production has shown that children encounter particular challenges when marking dative case (see, e.g., Ulrich et al., Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016). These challenges persist until school age. The most prevalent errors observed in production were overgeneralization of accusative case and the omission of determiners. In comprehension, children generally performed more accurately in acting out the appropriate action. If they acted out sentences incorrectly, they tended to show the action of a sentence with accusative case marking.
1.3. The relationship between comprehension and production in language acquisition research
It is generally argued that children must first understand a certain structure before they can use it productively, thus leading to the assumption that comprehension precedes production (Hendriks & Koster, Reference Hendriks and Koster2010). In contrast, other researchers have demonstrated a production/comprehension asymmetry, e.g., for Principle A and Principle B in Binding Theory (see, e.g., Bloom et al., Reference Bloom, Barss, Nicol and Conway1994; Hendriks & Spenader, Reference Hendriks and Spenader2006). Hendriks and Koster (Reference Hendriks and Koster2010) argue that the relationship between comprehension and production is less understood, as often only one modality is tested. It is crucial to acknowledge the interconnection between comprehension and production. Children not only process input to form knowledge of grammatical structures and word representations, but also possess the capacity to produce these structures and words. In addition to the observed asymmetry in comprehension and production, studies have identified instances of symmetry, such as the equal ability to comprehend and produce pronouns (Ruigendijk et al., Reference Ruigendijk, Friedmann, Novogrodsky and Balaban2010).
Given the established link between comprehension and production, and the ongoing discourse surrounding the identification of (a)symmetries in these domains, the present study aims to investigate children’s case marking skills in comprehension and production among the same participants.
1.4. Current study
The purpose of this study is to measure German-speaking preschoolers’ skills in accusative and dative case marking with local one-case and two-case prepositions in comprehension and production and to test whether age modulates their case marking skills. In addition, we will explore children’s non-target responses.
As outlined above, the majority of research on case marking in monolingual children focuses on children up to four years of age. Recent research has shown that children incorrectly mark dative case up to the age of nine (Ulrich et al., Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016). In our study, we focus on preschoolers between the ages of 4;5 and 6;5 to see how children older than four years understand and produce case. It is important to observe how preschoolers perform in tasks that involve case marking, as case marking is considered to be an essential component of language proficiency, particularly in the context of school-based learning (e.g., Deacon & Kirby, Reference Deacon and Kirby2004; Nation & Snowling, Reference Nation and Snowling2000). Another novelty is that we test a larger number of one-case and two-case prepositions (= six one-case and six two-case prepositions) compared to previous research.
We aim to answer the following questions:
-
1. Which case (accusative or dative) assigned by one-case and two-case prepositions is easier in comprehension and production for monolingual preschoolers?
-
2. Which type of preposition (one-case or two-case) is easier for monolingual preschoolers in comprehension and production?
-
3. Does the type of modality (comprehension or production) affect case marking with one-case and two-case prepositions in monolingual preschoolers?
-
4. What impact does age have on case marking skills in the production and comprehension of monolingual preschoolers?
-
5. What types of errors are attested in production?
We predict that children understand and mark accusative case more accurately than dative case (see Tracy, Reference Tracy1986; Turgay, Reference Turgay2011) and that it is easier for them to understand and produce case assigned by one-case prepositions than case assigned by two-case prepositions. Further, we expect an asymmetry between comprehension and production (with comprehension being more accurate than production). Concerning potential types of errors, we expect to find no case marking and overapplication of one of the cases, mainly accusative when dative case is required (Turgay, Reference Turgay2011).
2. Methods
2.1. Participants
Sixty-one monolingual children participated in this study. To obtain some background information on them and to ensure that there were no visual, hearing, or cognitive developmental impairments, all parents gave written consent and filled out a parental questionnaire. Using one standardized test and two other tasks (see Table 3) as inclusion and exclusion criteria, we assessed whether the participants know the prepositions in “into, in,” auf “onto, on,” unter “under,” vor “in front of,” hinter “behind,” neben “next to,” and zwischen “between” (PDSS, Kauschke & Siegmüller, Reference Kauschke and Siegmüller2012) and the gender of our test items and whether they are able to discriminate phonemes (PhoMo-Kids, Stadie & Schöppe, Reference Stadie and Schöppe2013). Children were excluded if they scored below the normal range in their age group in the phoneme discrimination task and below 50% correct in the tasks assessing knowledge of prepositions and gender. Two children were excluded, one due to these criteria and another one due to lack of attention during testing. A final sample of fifty-nine monolingual children was analysed (23 girls, M = 5;7, age range: 4;5–6;5, 5 children between 4;5 and 4;11, 38 children between 5;1 and 5;11, and 16 children between 6;0 and 6;5).
Table 3. Participant information and results of baseline assessments that were used as inclusion and exclusion criteria
1 Adapted subtest 9 from the PDSS (Kauschke & Siegmüller, Reference Kauschke and Siegmüller2012).
2 Subtest A1 from the PhoMo-Kids (Stadie & Schöppe, Reference Stadie and Schöppe2013).
3 Test of gender assignment to nouns, developed by the authors.
Besides the three tasks described above that served as inclusion and exclusion criteria, two additional tasks were administered: marking of accusative and dative case assigned by verbs and prepositions (ESGRAF 4–8, Motsch & Rietz, Reference Motsch and Rietz2019) and a test of the phonological working memory for non-words (SETK 3–5, Grimm, Reference Grimm2001). The responses given in the parental questionnaire indicated that eight children had articulation difficulties (shetism, sigmatism, /tr/ in contrast to /kr/), but given their accurate performance on the baseline assessments, they were not excluded from the study. Nine children were tested remotely during the COVID-19 lockdown. The other children were tested in their kindergartens in different parts of the state of Berlin. To ensure the same procedure for children tested remotely and in person, we used pre-recorded video adaptions and PowerPoint presentations of the standardized tests following the manuals. This study was approved by the Ethics Committee at Freie Universität Berlin (046/2021). More information on the participants and their results in our baseline assessments, which are described in the appendix (Appendix A), can be found in Table 3.
2.2. Experimental design
Three experimental factors were manipulated in a 2x2x2 design (see Table 4): word order (canonical versus non-canonical), preposition (one-case preposition versus two-case preposition), and case (accusative versus dative). Following a Latin square design, an item list was constructed that contains sentences involving six local one-case prepositions – three for dative (aus “out of,” von “from,” zu “towards”) and three for accusative (durch “through,” gegen “against,” um “around”) – and six local two-case prepositions (auf “on/onto,” hinter “behind,” in “in/into,” über “above/over,” unter “under,” vor “in front of”). The most frequent prepositions were chosen based on childLex (Heister et al., Reference Heister, Würzner, Bubenzer, Pohl, Hanneforth, Geyken and Kliegl2011). As subjects of the sentences, only animate nouns were selected (animals: der Vogel “theMASC bird,” der Käfer “theMASC beetle,” die Ente “theFEM duck,” die Eule “theFEM owl,” das Huhn “theNEU chicken,” das Küken “theNEU chick”), and as objects inanimate nouns (das Haus “theNEU house,” das Fenster “theNEU window,” das Tor “theNEU goal,” das Sofa “theNEU sofa,” die Kiste “theFEM crate,” die Box “theFEM box,” der Käfig “theMASC cage,” der Zaun “theMASC fence,” der Tisch “theMASC table,” der Reifen “theMASC hoop”). Three verbs are used that are compatible with one-case and two-case prepositions (fliegen “fly,” springen “jump,” laufen “run”). The list consists of 96 items: 48 with the canonical word order (Der Vogel fliegt über das Fenster. “The bird flies over the window.”) and 48 with the non-canonical word order (Über das Fenster fliegt der Vogel. “The bird flies over the window.”). These 96 sentences were divided into 48 items for production and 48 items for comprehension. All test sentences were assigned to four lists that were randomized across participants. Each participant was presented with a list of 24 target items (12 for production and 12 for comprehension) and 16 practice items (8 for production and 8 for comprehension). The distribution of nouns with all three genders was controlled across the verbs, and it was ensured that no differences in terms of frequency were attested between critical chunks of the stimuli (two-tailed t-test), which were the type of preposition, the PP-internal DP, and the case marking. Regarding frequency, no significant difference was found between one-case prepositions and two-case prepositions (t(81) = −0.41, p = 0.68) and between the PP-internal DPs in relation to the type of preposition – one-case versus two-case (t(75) = −1.80, p = 0.08) – and in relation to the case marking – accusative versus dative (t(130) = −1.28, p = 0.20).
Table 4. Experimental design
Note: ACC = accusative; DAT = dative.
2.3. Materials
To express the sentences’ events, short video clips were recorded with crafted animals and toys to display the inanimate references. Each visual scene was paired with an auditorily presented sentence, which was recorded with mildly child-directed clear speech by a trained female native speaker of German.
A visual illustration of one tested stimulus for one-case prepositions is provided in Figure 4.
Figure 4. Illustration of the critical video frame for the one-case preposition item Die Eule läuft zu der Kiste. “The owl is walking towards the box.” The white arrow illustrates the movement of the animal in the video.
In contrast to one-case prepositions, where children are required to acquire a single case assigned by a specific preposition and a corresponding conceptual representation, two-case prepositions necessitate the assimilation of not only the preposition’s meaning (spatial relation) but also the intricacies of the two case marking options and the linkage of two disparate conceptual representations to their respective case marking. The minimal pair conditions for sentences with a two-case preposition, as illustrated in Figure 5, imply that the sentences are distinguished solely by the case marking on the determiner. The first DP, the verb, the preposition, and the second noun are constant, the sole source of variation being the case marking on the determiner of the second DP. The different case markings lead to different conceptual representations.
Figure 5. Illustration of the critical video frames differentiating between the meanings of the two-case preposition items Das Küken läuft in die Box. “The chick is walking into the box.” (left-hand side) and Das Küken läuft in der Box. “The chick is walking in the box.” (right-hand side). The white arrows illustrate the movement of the animal in the video.
2.4. Procedure
The baseline measurements and testing of our stimuli in comprehension and production were spread over two weekly sessions, with a maximum of 40 minutes each. To counterbalance the completion of our production and comprehension tasks and to avoid priming effects of one task on the other, half of the children did the production task in session one and the comprehension task in session two, and the other half vice versa. Both sessions were recorded using OBS (Open Broadcaster Software, The OBS Project Contributors, 2022). The recordings ensured that the experimenter could rewatch the sessions if the need arose to examine the children’s mouthing (e.g., dem Tisch “theDAT.M table” and den Tisch “theACC.M table” only differ in one phoneme). In the kindergartens, the experimenter and the child were in a separate room, sitting at a table, with a laptop with the test materials on it placed in front of the child, together with a parent or an educator if the child so desired. It was ensured that the parents or educators did not intervene or answer for the child. All tests were presented to the children as games that the experimenter and a puppet, Mimi the snail, wanted to play with them.
To test the children’s production skills, the sentence completion task (Burgess & Shallice, Reference Burgess and Shallice1997) was used. At the beginning, the puppet introduced itself and stated that it needed help from the child (see the instructions in the appendix (Appendix B)). At least four practice videos were used to ensure that the child entered the spirit of the game by speaking for the puppet. Then, the 12 target videos were played. After watching a video (e.g., the sentences shown in Figure 5), the beginning of the sentence was auditorily presented as “The chick is walking in…” and the child was asked to complete the sentence. The child saw in the video that the chick is walking in the box (dative interpretation), and the puppet said “The chick is walking in…,” expecting the child to say “in the box.” The child’s answer (marking dative or accusative) shows whether they are able to use their case marking skills. Thus, the children were only asked to actively produce the minimal portion of the sentence that contains the critical information, not an entire sentence.
For the comprehension task, a sentence scene verification task was developed, which is based on the sentence picture verification task (Clark & Chase, Reference Clark and Chase1972). To ensure the same procedure for children tested remotely due to the pandemic and for children tested in person, we decided against the use of an act-out or pointing task. As it is difficult to depict the difference between the directional and the static meaning for two-case prepositions, we recorded videos to depict the actions clearly. To reduce the mental load for the children, we used a sentence scene verification task. This means that the child watched a short video clip and was then asked to say whether the sentence heard during the clip correctly describes the action depicted in the video. The testing started with a video of Mimi the snail, who introduced itself, explained the task to the children, and told them that it needed their help (see Appendix B). Afterwards, four practice videos were played to see whether the children understood the task. Then, the presentation of the target videos started. The experimenter again emphasized that it was important to help the puppet. In the videos, the accusative or dative interpretation of a sentence with a two-case preposition was depicted, and then, the puppet told the child what it had seen. The child then had to decide whether this was correct or not. For example, the video depicts the accusative interpretation of a sentence (e.g., that the chick is walking into the box as it is not yet in the box), but the puppet produces dative case when describing what has happened in the video (see Figure 5). As what the puppet said does not correspond to what has been shown in the video, the expected answer to the question “The chick is walking in the box. Is this correct?” is “No.” In the absence of a case-counterpart for one-case prepositions – it is not possible to construct two sentences that share the same DP1, verb, preposition, and DP2 but differ in case marking – lexical mismatch was chosen to elicit “incorrect” as an answer to counterbalance the occurrences of yes/no responses with one-case prepositions as well. In this context, the sentence in Figure 4 would be depicted in the video, but the puppet would instead say “The owl is walking towards the ladder. Is this correct?” expecting the child to reply “No.”
2.5. Data coding and preparation of target and non-target responses for analysis
All responses were coded as “1” when adult-like and “0” when non-adult-like. Indefinite forms were scored as correct if there was a clear case marking, as in gegen einen Tisch “against aACC table.” If a different noun with correct case marking was used, it was rated correctly as well, e.g., gegen den Zaun “against theACC fence” instead of gegen den Tisch “against theACC table,” regardless of whether it had the same gender as the target noun. Non-target responses (such as using the wrong case or preposition) were coded and grouped into different non-target response categories (see the subsection Non-target responses in production and description of categories used for their classification in the following section).
3. Results
In this section, we will describe the results of our production and comprehension tasks. In production, we provide a detailed analysis of accusative and dative case marking with one-case and two-case prepositions and the errors children made when incorrectly marking case. A detailed description of the accuracy rates of our comprehension task can be found in the appendix (see Appendix C).
3.1. Data analysis
As the binary dependent variable, we used the proportion of correct answers in our scene verification and sentence completion tasks that was calculated for each child. All responses entered a generalized linear mixed model (GLMM) in R (R Core Team, 2024, version 4.4.1) using the lme4 package (Bates et al., Reference Bates, Mächler, Bolker and Walker2015b, version 1.1.35.5) and the glmer function.
The initial model included three factors, with two levels each: case (accusative versus dative), preposition (one-case preposition versus two-case preposition), and modality (comprehension versus production). For all factors, the successive differences contrast coding was chosen. Next to the fixed effects, we specified subject and item as random effects, as well as age as centred continuous covariate. Following Bates et al. (Reference Bates, Kliegl, Vasishth and Baayen2015a), the initial model was specified, including all factors in the fixed and random effect structure, testing for interactions as well. Estimation problems prevented the fit of planned models with all fixed factors and all random intercepts and slopes. Non-converging and overspecified models were dealt with by progressively simplifying the random effect structure until convergence was reached by using a principal component analysis (PCA) utilizing the rePCA function (Bates et al., Reference Bates, Kliegl, Vasishth and Baayen2015a) and until we found the most complex one that is supported by the data and has the highest goodness of fit (see Table 5 for model specification).
Table 5. Model specification and output
Note: Significant z- and p-values are highlighted in boldface.
acc = accusative; dat = dative; PP = type of preposition; one = one-case preposition; two = two-case preposition; com = comprehension; prod = production.
The analysis revealed main effects of age, case, and preposition (see Table 5). This means older children performed more accurately than younger children, marking of accusative was more accurate than marking of dative case (accusative 78%, dative 56%, p < 0.001), and marking of case assigned by one-case prepositions was more accurate than marking of case assigned by two-case prepositions (one-case prepositions 82%, two-case prepositions 53%, p < 0.001).
The analysis revealed an interaction between case and modality, which reflects that marking of dative in production posed a greater challenge, in comparison to the comprehension task and as generally compared to understanding and producing accusative case.
In relation to the first and third research questions, Figure 6 illustrates, for both modalities, how accurately children performed on accusative and dative case assigned by one-case and two-case prepositions. It reflects the findings of the GLMM analysis that the understanding and producing of accusative case were more accurate than the understanding and producing of dative case and that the production of dative case assigned by two-case prepositions was the most difficult for children.
Figure 6. Mean accuracy of accusative and dative case marking with one-case and two-case prepositions in comprehension and production. Error bars denote +/− 2 standard error of the between-subject variation.
Referring to the second research question, Figure 7 shows how accurately children performed on the two types of prepositions in the comprehension and production tasks. Two-case prepositions posed greater challenges to children.
Figure 7. Mean accuracy of one-case and two-case prepositions in comprehension and production. Error bars denote +/− 2 standard error of the between-subject variation.
The fourth research question deals with a potential age effect. In addition to a main effect of age, the GLMM analysis revealed an interaction between age and preposition, with younger children performing less accurately in marking case assigned by one-case prepositions than older children, as illustrated in Figure 8.
Figure 8. Correlation plot showing the relation of age in months for each child with the mean accuracy in the two types of preposition conditions; left panel: one-case preposition condition, right panel: two-case preposition condition. Each point denotes one child plotted here with the values from the production and comprehension tasks. Note that both cases (accusative/dative) are included in each of the types of preposition conditions. The shades denote +/− 1 standard error based on single correlations and are not taken from the model output.
3.2. Non-target responses in production and description of categories used for their classification
To elucidate the fifth question, we evaluated children’s non-target responses in the production task, where participants produced in total 249 non-target responses (35%) out of a total number of 708 responses. There were 70 (of 249) non-target responses with one-case prepositions (28%) and 179 (of 249) non-target responses with two-case prepositions (72%). They marked accusative case incorrectly in 64 (of 249) cases (26%): 13 of these occurred after one-case prepositions (5%) and 51 after two-case prepositions (21%). They marked dative case incorrectly in 185 (of 249) cases (74%): 57 of these occurred after one-case prepositions (23%) and 128 after two-case prepositions (51%). Overall, the children produced many more non-target responses with dative case marking than with accusative case marking and with two-case prepositions than with one-case prepositions.
Non-target responses were coded according to Schmitz (Reference Schmitz, Hole, Meinunger and Abraham2006), who described six error type categories for dative case, and Scherger (Reference Scherger2015), who defined corresponding categories for accusative case. All non-target responses were coded according to five categories to address the following research questions: (i) Which case is overgeneralized (accusative instead of dative or dative instead of accusative)? (ii) Do we find determiner omissions? (iii) Do monolingual children struggle with choosing the correct gender and preposition?
Figure 9 depicts the distribution of target and non-target responses for each condition.
Figure 9. Proportions of target and non-target responses in the five different categories. Each panel shows the distribution of correct and incorrect responses in the four different conditions (one-case_acc = accusative with one-case prepositions, one-case_dat = dative with one-case prepositions, two-case_acc = accusative with two-case prepositions, two-case_dat = dative with two-case prepositions).
The children produced 459 correct responses and 249 non-target responses, resulting in 708 responses in total. For each of the four conditions, 177 responses were analysed. In the first condition, accusative marking with one-case prepositions (one-case_acc), the children produced 13 (of 177) non-target responses (7%). In the second condition, dative marking with one-case prepositions (one-case_dat), 57 (of 177) non-target responses (32%) were produced, 13 of those being overgeneralization of accusative case (7%) as well as using the wrong gender (7%). The third condition, accusative marking with two-case prepositions (two-case_acc), was characterized by 51 non-target responses (29%): 23 of those were overgeneralizing dative case (13%) and 20 were gender mistakes (11%). In the last condition, dative case marking with two-case prepositions (two-case_dat), 128 non-target responses were given (72%), with 62 occurrences of overgeneralizing accusative case (35%) as the most common non-target response type. Among all conditions focusing on the non-target responses, the most common pattern was to use accusative case where dative case was required (75 occurrences, 30%). There were 19 occurrences where the children overtly marked accusative case instead of dative but used the wrong gender (8%).
To summarize, in 94 instances (38%), the children marked accusative case instead of dative case. Other non-target response types were overgeneralization of dative case (24 occurrences, 10%) and using the wrong gender (22 occurrences, 9%) or preposition (18 occurrences, 7%). The most prevalent non-target responses were found in the fourth condition (two-case_dat), with the categories overgeneralization and non-informative being the most common. The category of non-informative responses encompasses ambiguous cases. These are attributable to the characteristics of the German language. The children could have marked either nominative or accusative case instead of dative case.
4. Discussion
Our study tested the understanding and production of accusative and dative case marking with one-case and two-case prepositions among monolingual German-speaking preschoolers. Case marking in German, particularly with prepositions that can assign more than one case, is challenging for children. As case marking makes a contribution to sentence and text comprehension and narrative skills, it is an important ability children need for school. So far, case marking in German has mostly been analysed with small sample sizes (e.g., Clahsen, Reference Clahsen1984) and with a focus on spontaneous speech (e.g., Schmitz, Reference Schmitz, Hole, Meinunger and Abraham2006; Tracy, Reference Tracy1986). Analyses under controlled conditions with larger sample sizes have been limited to a few prepositions and mostly production (e.g., Ulrich et al., Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016). In the current study, six one-case and six two-case prepositions in comprehension and production were tested in one experiment, controlling for age. The results reveal that dative case is in general more challenging than accusative case and that case assigned by two-case prepositions is more challenging than case assigned by one-case prepositions for both dative and accusative case. The research questions that were addressed at the beginning of the article were as follows:
-
1. Which case (accusative or dative) assigned by one-case and two-case prepositions is easier in comprehension and production for monolingual preschoolers?
-
2. Which type of preposition (one-case or two-case) is easier for monolingual preschoolers in comprehension and production?
-
3. Does the type of modality (comprehension or production) affect case marking with one-case and two-case prepositions in monolingual preschoolers?
-
4. What impact does age have on case marking skills in the production and comprehension of monolingual preschoolers?
-
5. What types of errors are attested in production?
We start by answering the first question, which is related to potential differences in accusative and dative case marking. Turgay (Reference Turgay2011) showed that children performed less accurately on dative case with both one-case and two-case prepositions. She concluded that there is no difference in case marking with one-case and two-case prepositions. In contrast to Turgay (Reference Turgay2011), in this study, the numbers of one-case and two-case prepositions were counterbalanced, and the verbs across different conditions were controlled. Thus, minimal pair sentences differing only in case marking were used. The results indicated that the participants performed less accurately on dative case independently of the type of preposition, which is in line with Turgay’s results. However, contrary to Turgay, the children in our sample struggled with understanding accusative and dative case marking with two-case prepositions. The accuracy rates for understanding dative case assigned by two-case prepositions were at chance level, while Turgay’s study documented few errors in this area.
Regarding dative case, we found that it poses challenges for children older than four years of age, which is in line with Ulrich et al. (Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016). Comparing the results of our stimuli to Ulrich et al.’s (Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016) and Ulrich’s (Reference Ulrich2017) results, we see that the children in our sample performed less accurately in marking dative case assigned by two-case prepositions than the children in their sample. This might be attributed to the fact that the only preposition tested in Ulrich et al.’s study is hinter “behind,” which might be acquired earlier than the other prepositions we also tested. Furthermore, in Ulrich et al.’s elicitation task, participants were explicitly instructed to indicate the location (= dative case) or destination (= accusative case) of the animals. If they incorrectly replied by using the wrong case, the researchers intervened to prompt the appropriate case. This intervention may have encouraged the use of the correct case in children’s spontaneous speech, a factor that was not present in our study.
Explanations for the acquisition order of cases vary from input frequency to morphological marking. Dative case occurs more frequently in PPs, making it potentially easier to learn dative assigned by a preposition than by a verb (Wegener, Reference Wegener1995). However, a frequency-based explanation does not capture the pattern of our results. Alternatively, an explanation based on morphological marking suggests that unmarked forms are learned earlier than marked ones (Fliedl, Reference Fliedl1999; Smits et al., Reference Smits, Mortelmans and Willems2020). Since nominative and accusative case are unmarked/less marked forms, they should be acquired first. Dative is strongly marked and therefore learned later. Our results clearly support this explanation as accusative case (accuracy of 78%) was generally more accurate than dative case (accuracy of 56%). Although the frequency-based explanation does not provide a satisfactory account of the results obtained, a contrastive and highly frequent input of accusative and dative case assigned by different types of prepositions may assist children in optimally learning from the input.
The second question focuses on the type of preposition. In contrast to Turgay (Reference Turgay2011), an analysis of the different conditions indicated that there are discrepancies between case marking with one-case and two-case prepositions. These differences can be explained by the process of feature unification as part of the PT and the lexical-functional grammar (see Baten & Willems, Reference Baten and Willems2012). According to the PT, features are processed within and across phrases. For one-case prepositions, children are required to process the information within a single phrase, which is called phrase-internal feature unification. In contrast, with two-case prepositions, processing occurs across phrases. The latter type of preposition is characterized by the existence of grammatical minimal pairs. This type is more challenging for children to acquire as they need to map the semantic properties of the preposition onto two distinct events that are distinguished solely by the corresponding case marking. In contrast, with one-case prepositions, the meaning of the individual preposition is acquired along with the only possible case marking, thus leading the children to understand and produce case assigned by one-case prepositions more accurately than case assigned by two-case prepositions. Combining the feature unification part of the PT and the lexical-functional grammar with the message level of Garret’s sentence processing account can explain the acquisition of case marking with one-case and two-case prepositions in monolingual children. That said, our goal to test a wider range of prepositions simultaneously has introduced a potentially confounding factor, which might weaken the sentence processing-based explanation just proposed. While the number of mono- and bisyllabic two-case prepositions could be controlled (three mono- and three bisyllabic two-case prepositions), five out of six one-case prepositions were monosyllabic, resulting in slightly shorter sentences than those containing two-case prepositions. Future research should compare one-case and two-case prepositions of similar length.
In this regard, it would be valuable to investigate the order of acquisition of individual prepositions. Given our research questions, we decided to manipulate the occurrences of one-case and two-case prepositions, as well as of accusative and dative case marking in both canonical and non-canonical word orders. Using this complex design, the occurrence of each preposition could only be counterbalanced across modality. Future research is needed to disentangle the order of acquisition for individual prepositions, using a design that allows independent testing of the same preposition in each modality.
Coming to the third question, which deals with the type of modality, the participants achieved comparable results in both modalities, with an accuracy rate of 70% in comprehension and 65% in production. The results align with those observed by Turgay (Reference Turgay2011), who found that children performed more accurately in the comprehension task. Despite a numerical difference between the two modalities in our data, the statistical analysis did not reveal any significant differences between the two modalities. This contradicts the argument that comprehension precedes production. The lack of asymmetry we found can be attributed to the equal number of prepositions that were evaluated in both the comprehension and production tasks. While Turgay’s study tested two prepositions in comprehension (in and on), we tested twelve prepositions in comprehension and production in our study, thus providing a more comprehensive insight into children’s proficiency in case marking. The differences could also be related to the chosen paradigm. Utilizing an alternative paradigm, such as an act-out task, could have yielded different results in the comprehension task.
The fourth research question addresses the role of age in children’s case marking skills. The finding that older children performed more accurately than younger ones is in line with Turgay (Reference Turgay2011) and Ulrich et al. (Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016). In addition, age modulates the accuracy of case marking with one-case prepositions in our data. For two-case prepositions, no significant differences were found. All children, independent of age, performed less accurately with two-case prepositions than with one-case prepositions. Therefore, we argue that two-case prepositions with their dual case marking options are harder to acquire than prepositions with a unique case. The development of case marking and specifically case assigned by two-case prepositions might be observed to evolve after five years of age. Furthermore, it is possible that the children in our sample might not yet have the conceptual representation of sentences with a static interpretation (= dative case marking). Therefore, the testing of older children and children in different age groups will potentially provide insights into the mastery of dative case.
In other languages with case marking, a different pattern emerges. For example, Turkish-speaking children acquire the Turkish case marking system by the age of three (e.g., Ketrez & Aksu-Koç, Reference Ketrez, Aksu-Koç, Stephany and Voeikova2009). Similarly, Russian children do not encounter difficulties in acquiring case marking. Research on German and Russian bilingual children has shown that case marking is a more readily available cue in Russian, enabling children to rely on it for sentence interpretation (e.g., Kempe & MacWhinney, Reference Kempe and MacWhinney1999). Similar to German, Icelandic has prepositions that can assign accusative and dative case, resulting in directional (= accusative) and static (= dative) interpretations (Dehé & Kupisch, Reference Dehé and Kupisch2022). Nicholas (Reference Nicholas2011) showed that Icelandic children either omit the preposition while using case marking or produce both the preposition and case marking. For German, the idiosyncrasies constitute a significant challenge for children in their acquisition of the case marking system. These cross-linguistic differences highlight that the ease of acquiring case marking varies across languages and is influenced by the specific grammatical features of each language.
In addressing the fifth question about types of errors, the analysis of non-target responses in production revealed that the most prevalent was the overgeneralization of accusative case. This observation is corroborated by the findings of Schmitz (Reference Schmitz, Hole, Meinunger and Abraham2006), Tracy (Reference Tracy1986), and Turgay (Reference Turgay2011). We noticed article omission in only a few instances, which is contrary to Meisel’s (Reference Meisel1986) findings, where it was the most common error type. A detailed investigation of the four conditions – accusative case marking with one-case prepositions, dative case marking with one-case prepositions, accusative case marking with two-case prepositions, and dative case marking with two-case prepositions – revealed that non-target response patterns varied by condition. While children did not encounter difficulties in producing accusative case with one-case prepositions, non-target response patterns differed in the three other categories. In the categories dative assigned by one-case prepositions and accusative assigned by two-case prepositions, the rates of overgeneralizations and gender errors were equal. Specifically, accusative case was overgeneralized in the condition dative assigned by one-case prepositions, while dative case was overgeneralized in the condition accusative assigned by two-case prepositions. In the condition dative assigned by two-case prepositions, the most common non-target response patterns involved overgeneralizing accusative case, along with non-informative responses, including ambiguous case marking (accusative or nominative). It has been observed that children generally encounter difficulties with dative case marking (see Ulrich et al., Reference Ulrich, Knickenberg, Penke, Lüdtke and Motsch2016). In addition, in the presence of two-case prepositions, an alternative option exists of employing a different case, resulting in a grammatically correct sentence. In non-informative responses, the children produced either accusative or nominative case instead of dative case. There are two possible explanations: they either utilize the less morphologically marked case to produce a sentence or do not possess the conceptual representation of the static interpretation.
In grammar acquisition, the presentation of alternative forms concurrently is a pedagogical strategy aiming to facilitate children’s awareness of linguistic complexity. Providing both options – the directional and the static interpretation – fosters a comprehensive understanding of both semantic concepts. Specifically, with regard to case marking with two-case prepositions, children must be presented with sentences that incorporate both directional and static interpretations. Through this approach, children can develop a nuanced understanding of when to employ accusative and dative case, thereby facilitating the development of both concepts.
The inherent characteristics of the German case marking system with no clear form-to-meaning mapping resulted in numerous responses that were deemed to be ambiguous, thereby influencing the analysis of non-target responses in the production task. One potential solution to this issue would be to limit the scope of the study to specific genders. Given that accusative and dative case marking has been tested in this study, this would mean limiting the testing to masculine nouns. However, it should be noted that feminine and neuter nouns have distinct forms for dative case. In order to avoid limiting the testing to specific genders, it is imperative to investigate the accuracy rates across all three genders. This would show whether the noun’s gender influences the rate of non-target responses.
5. Conclusion
The results presented in this paper show that two-case prepositions are indeed difficult for preschoolers. Age differences were found in the mastery of one-case prepositions. Older children performed significantly more accurately. This could lead to the conclusion that the acquisition of two-case prepositions may fully develop after children have mastered one-case prepositions completely, meaning that one-case prepositions with their single, unique case are understood and produced earlier than two-case prepositions.
We conclude that the acquisition of the German case marking system is a long process that is not completed by school entry. As dative case and two-case prepositions are particularly difficult for children and the mastery of marking the correct case is obligatory for subsequent language skills, training of these skills before school entry might be helpful for all children. The development of case marking skills can be enhanced by implicit training focusing on a contrastive and frequent input of accusative and dative case assigned by different prepositions. This focused stimulation can be incorporated into storytelling, reading stories, or singing songs with preschoolers, or it can be used to enrich caregivers’ language with accusative and dative case marking when it is appropriate to the situation (see, e.g., Egert et al., Reference Egert, Cordes and Hartig2022). Such focused stimulation has been shown to assist children in deducing grammatical information from the input (see, e.g., Bruinsma et al., Reference Bruinsma, Wijnen and Gerrits2020). In our particular case, it may facilitate the acquisition of the appropriate case marking in conjunction with the corresponding conceptual representation. With school-aged children, case marking with prepositions can be learned explicitly by integrating playful exercises into lessons. Existing card games that foster the acquisition of case marking with prepositions can be used for this purpose.
In order to obtain a more comprehensive understanding of children’s case comprehension and production of case marking with prepositions, it is imperative to extend the scope of the analysis to more prepositions. The present study has demonstrated that the implementation of a comprehensive evaluation encompassing six one-case and six two-case prepositions provides a nuanced and differentiated perspective on children’s aptitude in comprehending and producing accusative and dative case assigned by prepositions. Furthermore, the use of production and comprehension tasks offers a more refined perspective on their actual abilities to understand and produce the correct case.
Acknowledgments
We thank all the children, their families, and kindergartens who participated in the study and our student assistants who helped in recruiting the children. The project received funding from Freie Universität Berlin, which is gratefully appreciated.
Competing interests
The authors declare none.
Disclosure of use of AI tools
The authors declare that no AI tools were used during the preparation of this manuscript.
Appendix
Appendix A: Overview of the baseline assessments and additional tests used in the study
Note: DP = determiner phrase; PP = prepositional phrase.
Appendix B: Instructions given by the puppet used in the production and comprehension tasks
B.1. Instruction production Mimi the snail
Ich bin Mimi, die Schnecke. Ich habe Menschensprache gelernt, aber ich mache manchmal noch Fehler. Du siehst gleich ein Video. Schau es dir in Ruhe an. Dann fange ich an zu sprechen. Bitte hilf mir und sprich für mich weiter. “I am Mimi the snail. I learned human language but I sometimes make mistakes. You will watch a video now. Watch it carefully. Then I will start saying what I have seen. Please help me and continue speaking for me.”
B.2. Instruction comprehension Mimi the snail
Ich bin Mimi, die Schnecke. Ich habe Menschensprache gelernt, aber ich mache manchmal noch Fehler. Du siehst gleich ein Video. Schau es dir in Ruhe an. Dann sage ich, was ich gesehen habe. Hör genau zu und hilf mir. Sage mir dann: Passt das, was ich sage, ganz genau zu dem Video? “I am Mimi the snail. I learned human language but I sometimes make mistakes. You will watch a video now. Watch it carefully. Then I will tell you what I have seen. Listen to me carefully and help me. Tell me afterwards: Did my utterance exactly match the video?”
Appendix C: Non-target responses in comprehension
In our comprehension task, the children incorrectly judged the videos 214 times (out of a total number of 708 trials), with 58 occurrences in the one-case preposition condition (27%) and 156 occurrences in the two-case preposition condition (73%). The distribution of accusative and dative case is as follows: 90 occurrences of accusative case (42%) and 124 occurrences of dative case (58%). A close examination of the two cases and the two types of prepositions reveals the following non-target response rates for the four conditions: 21 non-target responses related to accusative assigned by one-case prepositions (12%), 37 related to dative assigned by one-case prepositions (21%), 69 non-target responses with accusative assigned by two-case prepositions (39%), and 87 with dative assigned by two-case prepositions (49%) (see Figure C1).
Figure C1. Proportions of correct and incorrect responses. Each panel shows the distribution of correct and incorrect responses in the four different conditions (one-case_acc = accusative with one-case prepositions, one-case_dat = dative with one-case prepositions, two-case_acc = accusative with two-case prepositions, two-case_dat = dative with two-case prepositions).
Open access
