Sign and indexicality: a case study of enhancing alignment of situation models in SCWT

2.1 The Xu-argument based SCWT as a proficiency test

Xu-argument, a newly emerging type of view on second language acquisition, claims that language is acquired through xu (续 in Chinese), a Chinese character with a meaning of completion, extension, and creation (Wang 2016). The Xu-oriented research has tested the principles of Xu-argument and the high efficiency in language learning achieved by Xu. It has attracted increasing research attention, as witnessed by numerous articles published in Chinese domestic journals (Q. Wang 2019; Wang 2014; Wang and Hong 2021; Wang and Qi 2013; Wang et al. 2000; Yang 2018; Yang and Lu 2021) and international journals (Chen and Zhang 2022; Wang and Wang 2015; Zhang 2017; Zhang and Zhang 2021). Those studies have demonstrated that continuation tasks can facilitate language learning in various aspects (Jiang and Chen 2015; Wang et al. 2021; Xin 2017; Ye et al. 2021; Zhang et al. 2022).

Some studies have been conducted to verify the validity of SCWT as a proficiency test. Wang and Qi (2013) conducted the first empirical study to examine if SCWT is suitable for the use in L2 proficiency test. Data were collected from a sample of 334 grade 3 senior high school L2 learners of English in Guangdong Province. The findings lead to the conviction that the incorporation of SCWT into the NEMT writing section can contribute to the assessment objective of promoting language learning. Hu (2021) conducted a content validity analysis of SCWT in the NMET (Zhejiang Province and Shandong Province) in 2020. The findings indicate that SCWT is not only in line with test-takers’ English proficiency level in China’s EFL context, but also in line with the requirements of General High School English Curriculum Standards 2017.

A study that substantially promotes the understanding of current rate quality of SCWT is Chen and Zhang (2020), which examines the effectiveness of measures of enhancing the rate quality of SCWT in the NMET involving 1,607 grade 3 students in senior high schools. Results indicate that the current NMET rating procedures are suitable for ensuring the rating quality of SCWT.

SCWT has been incorporated into China’s NMET in Zhejiang Province since 2016, Shandong Province since 2020, and Jiangsu Province since 2021. However, the application of SCWT in the NMET context is still in its infancy. Both senior high school students and teachers struggle with many adjustment problems. The potential might be overwhelmed by the requirement of demonstrating one’s writing ability within the testing context (Cohen and Upton 2007). Therefore, more empirical studies should be conducted to accumulate evidence with respect to test-taking response process to promote its assessment use (Zumbo and Hubley 2017).

2.2 Alignment of situation models and meaning-related problems

A phenomenon of emerging interest within language learning is that of alignment (Atkinson et al. 2007; Costa et al. 2008; Pickering and Garrod 2004). Pickering and Garrod (2004) argue, from a psycholinguistic perspective, that alignment is the basis of successful communication in dialogue. Atkinson et al. (2007), from a social cognitive perspective, hold a holistic view, regarding alignment as a default process of continually aligning oneself with one’s sociocognitive environment. In keeping with the conception developed by Pickering and Garrod (2004) and Atkinson et al. (2007), Wang (2012) claims that like interpersonal communication, learner-text interaction involves alignment as well. Wang (2014) argues that the role of alignment in L2 learning occurs at various levels of representation, including linguistic representations (phonetic, phonological, lexical, syntactic, and semantic) and situation models.

Alignment of situation models forms the basis of successful communication (Pickering and Garrod 2004). A situation model is a multidimensional representation of the situation under discussion (Sanford and Garrod 1981), involving five dimensions of space, time, causation, intentionality, and protagonist (Zwaan and Radvansky 1998). The majority of studies on five key situational dimensions have been conducted in isolation (Zwaan and Radvansky 1998). Despite a few studies that have investigated the interrelatedness of the situational dimensions (i.e., Taylor and Tversky 1997; Zwaan et al. 1995), little attention has been directed to integrate key dimensions to enhance alignment of situation models. One study (Zhao 2021), in particular, investigated the effect of situational alignment on producing less meaning-related errors in L2 writing, and confirmed its strength. Nonetheless, this study is exclusively concerned with English majors at university, and employed true or false reading comprehension task concerning situational dimensions to enhance alignment. The strategy is applicable exclusively in empirical experiment condition, rather than within the testing context.

The alignment of situation models in SCWT may be more controlled and strategic than automatic due to a lack of gesture, posture, facial expression, and other environmental affordances available in dialogue. Learners have to consciously align with the situation model built on the input text by going through the following procedure: “first, they need to construct a situation model for comprehension based on the text. Secondly, they need to draw on the situation model so constructed to develop ideas for writing. Thirdly, they need to make sure that their writing coheres or aligns with the situation model they have constructed” (Wang and Wang 2015, p. 507). The procedure provides a road map, but doesn’t undoubtedly ensure correct and appropriate use of L2 and coherence with the input text. For instance, both senior high school students and teachers spare all efforts on reciting chunks and sentences describing smile, tears, joy, sorrow, horror, courage, etc. But most test-takers still encounter setbacks and difficulties in improving meaning-related expressions, which result in poor exam results, no constructive feedback, learning plateaus, and learning demotivation. While senior high school English teachers complain about the new writing task, some of them even tell students that SCWT will last no longer than three years. Little attention has been directed to completing the story in a coherent and logical way. The alignment of situation models cannot be enhanced through the identification of five key situational dimensions in isolation or through the interference of true or false reading comprehension tasks within the testing context. Hence, potentially costly strategies are required when automatic mechanism cannot produce alignment (Costa et al. 2008; Pickering and Garrod 2004, 2006). The present study just aims so.

2.3 Pan-indexicality model as a solution

One of the key problems of SCWT is that test-takers are not competent in ensuring their written productions to cohere or align well with situation models. Indexicality, one aspect of sign, focuses on sequential or causal relations that feature certain logical inferential properties (J. Wang 2019). Therefore, the present study adopts the framework of a pan-indexicality model to investigate the factors that contribute to the development of situational alignment in SCWT in the NMET context.

Pierce’s trichotomy of symbol, icon and index has been an effective tool for analyzing verbal and non-verbal signs. According to Peirce (1955), index and indexicality are two distinct concepts, the former being associated with a certain type of linguistic sign, while the latter relating to a property of reference that is represented in the former. However, some scholars (e.g., Burge 1982; Doane 2007; J. Wang 2019) have realized that the definition of index and indexicality needs to be adjusted so as to be at the same analytic level as symbol and icon.

Based on the intrinsic connection between sign, index and indexicality in Peirce’s semiotics, J. Wang (2019) proposes his definition of indexicality in terms of pan-index and pan-indexicality. All words are taken as pan-index and contain the essential property of pan-indexicality. When all words are taken as pan-index, a pan-indexicality model is thus proposed to interpret linguistic signs from a brand-new perspective. Figure 1 is a description of the pan-indexical process that affects meaning comprehension.

Figure 1:

The pan-indexical process of a linguistic sign (J. Wang 2019:57).

Figure 2:

Kang’s three-phase process of situational alignment development.

Figure 3:

The pan-indexical process of “the paper bag”.

According to J. Wang (2019), the pan-indexical process interprets a linguistic sign by following Peirce’s classic sign process of sign, object and interpretant, and incorporating the idea of taking all signs as having the basic property of indexicality. A linguistic sign is basically associated with a certain object that waits to be interpreted by an interpretant. The interpretant is classified into interpretantⁱ and interpretantⁱⁱ. Interpretantⁱ, which is concerned with shared information understood by both the addresser and the addressee, assures the accurate reading of what the object refers to. Interpretantⁱⁱ, which is about the addressee’s personal encyclopedic knowledge, perception of the contextual information, and personal emotions in the process of the actual communication, influences the final reading of the linguistic sign. In this model, a linguistic sign reveals formal, conventional and overall sense of the object. Formal sense is understood almost automatically, such as the acoustic form of hiccup which would evoke a sharp, usually repeated, sound made in the throat. The conventional sense, presented as an entry in a dictionary, is a general agreement reached by both the addresser and the addressee. The overall sense is solely associated with interpretentⁱⁱ or the addressee’s personal encyclopedic knowledge, perception of contextual information and personal emotions.

In the pan-indexicality model, a linguistic sign is regarded as an index that guides the addressee to understand the meaning of an object at the overall sense level. In SCWT, under the framework of a pan-indexicality model, a linguistic sign or an object encoded and stored in a causal chain of a narrative story might afford opportunities for test-takers to decode the meaning of a linguistic sign and infer the intentionality, which would contribute to the development of alignment of situation models.

All in all, theoretical conceptions and empirical evidences presented above illustrate that the application of SCWT as a proficiency test in the NMET context can generate positive washback, and alignment of situation models form the basis of successful learner-text interaction in SCWT. However, little attempt has been made to explore test-takers’ perspectives regarding how to enhance alignment of situation models. Therefore, the present study, by adopting the framework of a pan-indexicality model distinguished from other alignment-oriented studies so far, aims to investigate the factors that contribute to the development of situational alignment in SCWT in the NMET context.

3.1 Participant

The case selection strategy adopted in this study is intensity sampling, which refers to “a selection of information-rich case but not extreme” (Duff 2011, p. 106). One participant, Kang, took part in this study. The first author (Zhu) knew her well prior to this study. She coached Kang in a tutorial session for nine years. The participant was a grade 3 senior high school student in Suzhou, Jiangsu Province, when this study was carried out. She was willing to share her learning experience of SCWT with the researchers.

Kang’s English proficiency level was among top 5% of her class. At the beginning of the third year in senior high school, she usually got full marks in sections of listening and reading comprehension, 12–13 out of 15 as full marks in the section of English letter writing, but only got 17–18 out of 25 as full marks in SCWT section. She had sufficient confidence in her performance in English tests, but she felt stressed at times because SCWT occupied most of her time spending on English learning, yet yielded the least progress. In her written production of SCWT, there was no shortage of vocabulary or chunks. What she concerned about was how to enhance the alignment of situation models, to reduce meaning-related errors, and ultimately to achieve a higher score of above 20 marks.

3.2 Research questions

The present study, by adopting a qualitative approach, aims to address the following questions:

1. How do grade 3 senior high school students in Jiangsu Province enhance the alignment of situation models in SCWT?

2. What factors influence the development of situational alignment in SCWT?

3.3 Instruments

Research instruments, including face-to-face interviews, face-to-face conversations, and email exchanges, were elaborately designed based on Zwaan and Radvansky’s (1998) five dimensions of situation models, and J. Wang’s (2019) pan-indexicality model.

3.4 Data collection and data analysis

Three semi-structured individual face-to-face interviews in Chinese and each lasting around 2 h, were conducted with the participant to elicit in-depth data. Prior to the interviews, she was guided to sign a consent form acknowledging that 1) she participated in the study voluntarily; 2) she had the right to withdraw from the study at any time of the process; and 3) all interviews were video-taped for research purpose only.

The present study lasted over one year. The procedure is as follows: The first round of interviews was carried out at the beginning of the academic year. The interviews focused on the identification of five dimensions of situation models. The second round of interviews, conducted at the end of term 1, provided the researchers with an access to the latest updates on her development and problems regarding the alignment of situation models. The third round of interviews, conducted at the end of term 2, mainly focused on how she integrated key dimensions of situation models under the framework of a pan-indexicality model and enhanced the alignment intensity in SCWT. During the interviews, open-ended questions were proposed to allow the participant to “create the options for responding … without being forced into research responsibilities” (Creswell and Plano Clark 2006, p. 32).

Data were collected through face-to-face conversations and e-mail exchanges. Information collected from the above sources enriched interview data by depicting a fuller picture of the participant’s development in the alignment of situation models.

Qualitative data analysis employs an “inductive investigative strategy” (Merriam and Tisdell 2009, p. 39). The interviews were first transcribed, coded, and compared with the data collected from supplementary sources to ensure consistency. The thematic structure of the case was then investigated for the theme, which was discussed with reference to research questions.

5.1 Development of situational alignment in SCWT

The development of situational alignment of the participant can be divided into three phases (see Figure 2):

Equipped with the knowledge of five dimensions of situation models, the participant was able to capitalize on protagonist-related, temporal and spatial dimensions to complete a story, but with some meaning-related errors in her written production in the first phase. The action that the Mexican family took the money as a reward and gave back a souvenir and a thank-you card, leads to the realization of a theme of offering help and expecting a reward in return, which turns out to be the opposite or at least not in line with the core values.

In the second phase, the participant validated causal relations (Zwaan and Radvansky 1998) between events described across paragraphs by reading back of the first sentence of the last two paragraphs. She constructed a channel of developing her situational alignment by forming causal connection between events described across paragraphs, with the absence of causal connectives.

In the third phase, the participant drew motivational inferences through encyclopedic knowledge, contextual information and personal emotions. Failed goal information is more available than completed goal information (Lutz and Radvansky 1997) presumebly because completed goal information is often part of the causal chain of a narrative story that connects several ideas together across paragraphs. Moreover, goal information is seldom stated explicitly in the input text and has to be inferred. Indexation may be needed in expert-novice interactions (e.g., Isaacs and Clark 1987). The process of writing the last two paragraphs of the story in SCWT involves interactions between the author and test-takers, which is similar to that between expert and novice. Kang constructed the overall sense of the paper bag by integrating her encyclopedic knowledge of a paper bag holding packaged food, her perception of contextual information of not having lunch yet, and personal emotion of a poor but hardworking and warmhearted family offering help without expecting any reward in return.

The above three phases showed that the participant developed her alignment of situation models in SCWT through identifying and integrating temporal, spatial, causal, motivational, and protagonist-related dimensions. This finding is different from Zhao’s (2021) finding that the alignment of situation models could be enhanced through True or False reading comprehension task.

5.2 Factors enhancing situational alignment in SCWT

That the participant was able to enhance her capacity to construct aligned situation models is to a large extent attributable to the strategy of identifying key dimensions of situation models and integrating these dimensions in a particular way.

As for identifying and integrating key dimensions of situation models, the causal and motivational dimensions form the backbone of situational alignment (Zwaan and Radvansky 1998). The temporal information is essential to the construction of situation models because it identifies causal and motivational links between events. Test-takers need to figure out the temporal order of events before generating a causal relation. Spatial information is often encoded explicitly in the input text, though not as closely intertwined as temporal information with causal and motivational dimensions. Finally, protagonist-related dimension is a vital component of situational alignment because it includes more information about protagonists than just their names. Most narrative stories describe the goals and actions of main protagonists. In addition, objects, which are often stored and encoded presumably depending on their causal connections, can also function as a vital element of situation models.

To avoid having the situational alignment excessively complex, “a token often contains a pointer that refers to more generalized information about an entity” (Zwaan and Radvansky 1998, p. 180). The data clearly revealed how a pan-indexicality model (J. Wang 2019) afforded opportunities for test-takers’ efficient integration of key dimensions of situation models. A test-taker’s personal encyclopedic knowledge, perception of the contextual information and personal emotions interact with each other, with some of them being more active while others less so in some cases, constructing the overall sense and exerting influence on what the linguistic sign ultimately means. A description of the pan-indexical process of “the paper bag” can be shown as follows (see Figure 3):

While the non-linguistic feature of “The paper bag” carries the cultural connotation of takeaway or fast food, the present study analyzes it mainly from the perspective of the linguistic feature. “The paper bag”, a crucial element, was taken as a linguistic sign to guide the participant to find out what objects the Mexican family placed in it. At the formal sense level, “the paper bag” refers to a bag made from paper. At the conventional sense level, it refers to a carrier bag made from paper, for holding purchases. At the above two levels, the participant obtained the information shared by both the author and the reader. Then move forward to the overall sense level. Kang’s encyclopedic knowledge of paper bag used for holding food in the Western countries makes the contextual information of “the girl asked if I had lunch” integrated with the key word “paper bag” by way of a causal connection. Kang’s personal emotion drove her to bring the story to an end in a way that the $20 bill would be returned by the poor but hardworking and warmhearted Mexican family. The object, the $20 bill, is part of a causal chain of the story and generates motivational influence to the theme of helping others without expecting any reward in return.

Overall, the data clearly reveal how a pan-indexicality model afforded an opportunity for the participant to use her encyclopedic knowledge, perception of contextual information and personal emotions to decode the meaning of a linguistic sign at the overall sense level accurately, which contributed greatly to the identification and integration of key dimensions of situation models. As a result, the participant was able to enhance her alignment of situation models in SCWT. This finding is in line with J. Wang’s (2019) assertion that knowledge owned solely by the addressee plays a key role in understanding the meaning of a linguistic sign at the overall sense level. It is also in line with Pickering and Garrod’s (2004, 2006 assertion that strategies are required when automatic mechanism cannot produce alignment.

The above findings suggest that the strategy of identifying and integrating situational dimensions by exploring the meaning of a linguistic sign at the overall sense level in a pan-indexicality model, from the perspective of encyclopedic knowledge, contextual information, and personal emotions, could enable test-takers to enhance their alignment of situation models in SCWT. It should be noted that despite the contribution of a pan-indexicality model in SCWT, it remains unknown how the model influences situational alignment in the comprehension and production of other types of texts, like expository and argumentative texts.