This compact book by Bert Cappelle in the Elements in Construction Grammar series (60 content pages) is a significant contribution to the theory of Construction Grammar (CxG). It poses a highly relevant and, dare I say, provocative question: Can Construction Grammar be proven wrong? As such, it can be considered a direct follow-up to an equally provocative question that Hoffmann (Reference Hoffmann2020) posed earlier: What would it take for us to abandon Construction Grammar?

The point of departure is Karl Popper’s (Reference Popper1963, Reference Popper, Lakatos and Musgrave1970, Reference Popper2002[1959[1935]]) approach to science, advocating for ‘falsifiability as a crucial property for science’ (p. 15). Applied to CxG, this translates to the question posed in the book’s abstract: Can every postulated construction … be empirically demonstrated not to exist? (my emphasis). Step by step, through six well-structured chapters, Cappelle takes us closer to answering this question, eventually arriving at a conclusion that may surprise some readers.

In essence, as stated in the abstract, this book is a quest for finding ‘a crucial test for CxG itself, one that would falsify it as a theory’. After exploring various possibilities, Cappelle concludes that such a test would have to prove that speakers have stored ‘pure-form knowledge’, i.e. forms without meaning. This translates to ‘find[ing] structural patterns that the community of [Construction Grammarians] can all agree on are not associated with some kind of meaning’ (p. 56, my emphasis). If such a ‘meaningless’ formal regularity could be found, it would contradict one of CxG’s most fundamental principles: ‘[a]ll of linguistic knowledge is a network of form–meaning pairs – constructions – and nothing else in addition’ (Hilpert Reference Hilpert2021: 6). Clearly, Cappelle’s aim is to encourage adherents of CxG to take a step back and (re-)evaluate what they are doing. This renders the book a valuable and fitting contribution to the discipline.

In the opening chapter, ‘Introduction: The stakes’ (pp. 1–7), Bert Cappelle establishes the essential context for investigating CxG as a significant linguistic theory. He notes the high regard in which CxG is held within the field of linguistics, emphasizing its inclusive approach that treats varied linguistic units as constructions. Cappelle also addresses – in my opinion very valid – concerns about complacency and dogmatism – such as ‘CxG is often reduced … to a few almost inspirational quotes’ (p. 1) – which he asserts are prevalent among proponents of CxG. In this context, he reminds us that ‘just because a linguist can detect a pattern, this doesn’t mean that the human mind belonging to an individual speaker also “sees” this pattern’ (p. 3), one of several calls for self-reflection. Cappelle also stresses that a critical examination is crucial for the theory’s vitality and warns of the dangers associated with CxG appearing to be ‘more a collection of unchallengeable assumptions’ (p. 2).

Cappelle introduces the notion of falsifiability, drawn from the philosophy of science, with a focus on Popper’s work. Popper’s approach holds that for a theory to be considered scientific, it must be testable and open to refutation. Applied to the theory of CxG, this translates to the primary question posed in the book’s abstract: Can every postulated construction within CxG be empirically demonstrated not to exist? This sets the scene for the ensuing exploration of CxG’s theoretical claims and its potential shortcomings.

Importantly, Cappelle notes that while CxG focuses on the mental representation of language in a way that may seem resistant to falsification, it is still vital to scrutinize the theory. By establishing the stakes of the inquiry early on, Cappelle invites a robust discussion that balances the accomplishments of CxG with a critical evaluation of its foundational claims.

Chapter 2, entitled ‘Can’t touch this: Does Construction Grammar have an attitude problem?’ (pp. 7–14), comprises two subchapters and covers the popularity that CxG has come to enjoy (section 2.1 ‘Being on a roll’) and the dangers associated with such success (section 2.2 ‘Dangers’). This chapter thus addresses a sociocultural aspect of CxG, which Cappelle calls an ‘attitude problem’ (p. 7) among many of its practitioners. He argues that the community of Construction Grammarians exhibits a certain level of insularity that leads to marginalizing opposing theoretical perspectives. This includes, but is not limited to, colleagues from the generative tradition. In addition to the ‘echo chamber’ of CxG, Cappelle maintains that there are also ‘echo corners’ (p. 14), warning that this insularity can hinder the progression and refinement of CxG by inhibiting critical discourse.

Drawing on the historical context of linguistic theory, he observes that many contemporary linguists are less inclined to engage with debates that have shaped the field. This tendency could lead to a stagnation in theoretical advancement. Furthermore, Cappelle critically examines the functional-cognitive approach of CxG, which positions itself in opposition to the generative framework. He advocates for more openness to findings from other approaches, including insights from neuroscience, which may challenge entrenched beliefs about language capabilities.

In essence, in this chapter Cappelle calls for a more collaborative spirit within the linguistic community, not shying away from admitting that generative linguistics ‘may be right about one thing: humans have a special knack for language’ (p. 14). He posits that more honest discussions of theoretical frameworks could lead to more nuanced understandings of language and enhance the scientific credibility of CxG.

Chapter 3, ‘Falsificationism: A still-influential approach to scientific inquiry’ (pp. 14–21), comprises three subchapters in which Cappelle shifts the focus to the philosophy of science, particularly the concept of falsificationism. He provides an introduction to Popper’s framework (section 3.1), which emphasizes that a scientific theory must be able to make predictions that can be empirically tested and, thus, potentially disproven. Cappelle provides good arguments for ‘relaxing’ Popper’s ‘dogmatic falsificationism’ (section 3.2), simply because it is sometimes downright ‘unrealistic’ (p. 17), and acknowledges that the falsification principle is therefore ‘just an ideal’ (p. 21).

In substance, in chapter 3, Popper’s approach serves as a critical lens through which Cappelle evaluates CxG. It includes a review of how falsificationism has shaped scientific inquiry and why it remains a crucial criterion for evaluating theories, including linguistic ones. Cappelle theorizes that CxG can strengthen its scientific grounding by framing its hypotheses in a manner that lends itself to empirical testing. He examines what constitutes a falsifiable hypothesis within CxG and proposes ways to demonstrate the potential nonexistence of specific constructions.

The chapter also discusses the relationship between linguistic theory and empirical data, pointing out the necessity for theories to accommodate linguistic variability and be adaptable based on new evidence. It emphasizes the importance of seeing falsification as an opportunity for growth rather than merely a challenge to be circumvented.

In the fourth chapter, entitled ‘The particular but generalizable problem posed by particle verbs’ (pp. 21–38), Cappelle explores the complexities associated with the ‘transitive verb-particle pattern’ (p. 3), presenting a case study that highlights issues with a CxG-based analysis. The choice of particle verbs for this case study is not coincidental, since they have been the object of Cappelle’s interest for well over two decades (cf. e.g. Cappelle Reference Cappelle, Dehé, Jackendoff, McIntyre and Urban2002, Reference Cappelle2006, Reference Cappelle, Bergs and Diewald2009, Reference Cappelle and Aronoff2023). Particle verbs, also referred to as ‘phrasal verbs’ (p. 21), are combinations of verbs with prepositions or adverbs and pose distinctive challenges that illuminate theoretical problems concerning constructional schemas. In particular, the question is whether to treat them as words and, if so, how to account for the fact that they can be split up, leading to at least two ‘allostructions’ (p. 22). This, in turn, leads to the question whether (1) there is a ‘general construction hovering over the … specific [allostructions]’ (p. 22) or whether (2) it is sufficient to posit horizontal links between these allostructions, without a ‘mother’ construction. The former option is referred to as a ‘layered’ representation, the latter as a ‘flat’ representation (p. 28). The issue here is that a ‘mother node’ appears to be superfluous (‘no transitive verb-particle combination found in usage can be licensed by the mother node without also being licensed by either of the two daughters’, p. 23) and therefore ‘no data can ever be adduced to falsify its existence’ (Hoffmann Reference Hoffmann2020: 150). In other words, if this question cannot be resolved using empirical data, this would translate to an unfalsifiable claim.

In the subchapters that follow, Cappelle attempts to resolve this question, discussing ‘three kinds of potential evidence’ (p. 23) that might be adduced: psycholinguistic evidence (section 4.1), computational learning (section 4.2) and introspective linguistic analysis (section 4.3). After finding the first two to be insufficient, he settles on the latter, admitting that it ‘will not count as strong empirical evidence’ (p. 28). After extensive deliberation, Cappelle arrives at the interesting conclusion that the layered and flat representations are, in fact, just ‘notational variants’ (p. 32). Ultimately, he concludes that the ‘contentious proposal’ at hand ‘turns out “testable”’ (p. 38), resolving the issue of unfalsifiability in this particular case, before asking: ‘But what about the theory itself?’ (p. 38).

In chapter 5, ‘How Construction Grammar could play the science game fairly’ (pp. 38–59), Cappelle turns to the question of how the theory of CxG (as opposed to ‘lower-tier’ issues such as the transitive verb-particle construction discussed in chapter 4) might be falsified. He states three rules that CxG should play by if it is to be taken seriously (p. 41) and then, following these rules, embarks on the quest for finding a basic claim of CxG that is falsifiable (sections 5.1–5.3). After having stated the object and objective (section 5.1), Cappelle critically reviews some auxiliary assumptions of CxG (section 5.2) ‘without whose acceptance [the claim that all of our knowledge of language is made up of constructions] becomes less intuitively correct and thus more vulnerable’ (p. 43). Each of these assumptions – domain-generality (section 5.2.1), the syntax–lexicon continuum (section 5.2.2) and isomorphism (section 5.2.3) – is convincingly argued to have issues, although Cappelle is able to show that accepting these issues is ‘not crucial for CxG to survive’ (p. 52).

This brings us to the most important section of the entire book, which will offer an answer to its eponymous question: ‘Formulate a crucial test’ (section 5.3, p. 52) to prove CxG wrong as an entire theory. Cappelle suggests the following:

In the following sections, Cappelle explores various possibilities of testing for such meaningless patterns, including transitive and gapping (section 5.3.2), both of which he discards as unsuitable for this purpose. Rather, our best bet, he argues, are ‘constructed constructions’ (section 5.3.3, p. 56). To this end, he suggests an ‘imaginary syntactic judgment experiment’ to investigate the ‘main-verb-before-complement(s) rule’ (p. 57). If such an experiment were to show that ‘language users have access to knowledge about their language that is purely structural (e.g. “in VPs, the main verb appears before any complements”)’ (p. 57), Cappelle argues, it could be shown that CxG is indeed falsifiable. In other words, the question posed in the book’s title can be answered in the affirmative.

On a final note, Cappelle invokes ‘parallel architecture (PA)’, which ‘admits the possibility of schemas/constructions that do not involve semantics’ (p. 58) to soften the blow a little in case his experiment outlined above were to show that speakers indeed possess purely structural linguistic knowledge. This conciliatory tone is carried over into the final chapter.

In the concluding chapter, entitled ‘Keep calm and constructi-con’ (pp. 59–60), Cappelle synthesizes the key themes discussed throughout the book and reflects on the future of CxG. Despite the challenges and critiques that CxG faces, he asserts that it holds valuable potential for advancing linguistic understanding. Cappelle encourages CxG theorists to look forward and not be afraid of re-evaluating and amending assumptions they have held dear, for example by ‘admitting functionally void schemas’ (p. 59).

In this chapter, Cappelle thus reiterates the importance of continued critical inquiry, acknowledging that, while CxG may not achieve a perfect theoretical status, its bold conjectures about the nature of language serve as a stimulus for further exploration. Despite acknowledging that CxG ‘quite likely will be or already has been falsified’ (p. 60), he ends on a thoroughly optimistic note about the capacity of CxG to engender rich debates and drive forward linguistic research, closing with the observation that, while ‘[t]he theory may be somewhat wrong, … we are clearly doing something right’ (p. 60).

Throughout this monograph Can Construction Grammar Be Proven Wrong?, Bert Cappelle engages critically and sometimes ruthlessly with foundational aspects of CxG, arguing for a rigorous scientific approach that emphasizes falsifiability, empirical testing and openness to alternative perspectives. While he acknowledges the successes and contributions of CxG, he also highlights the importance of addressing its limitations and fostering collaborative discussions with diverse linguistic frameworks. The book serves as a call to action, if not a wake-up call, for Construction Grammarians, encouraging them to scrutinize and refine their theoretical commitments while embracing the complexities of language. In doing so, it stands as a thoughtful and thought-provoking contribution to ongoing discussions within the field of CxG. It is an honest, constructive and optimistic critique, offering guidelines for how the theory can further mature and enhance its standing within the overall scientific framework.

Cappelle reminds us that CxG must embrace transparency in its theoretical foundations and be willing to engage with empirical data in a manner that aligns with scientific inquiry. He proposes specific strategies for ensuring that claims made within CxG are falsifiable, suggesting that, in line with Popper’s approach, linguists define clear operational parameters around their hypotheses. In this book, Cappelle highlights the need for consistency between theoretical predictions and actual language use, advocating for an approach which does not shy away from testing the most fundamental hypotheses of CxG (what he refers to as the ‘core’). By implementing these principles, he believes that CxG, which is ‘not just a theory but a veritable program’ (p. 60), can solidify its credibility. As such, the book’s primary target audience is clearly the community of Construction Grammarians, but it should also be of interest to adherents of other approaches.

By advocating for an adaptive and open-minded approach within the CxG community, Cappelle leaves the audience with a sense that linguistic theory in general and CxG in particular is a dynamic field that can only benefit from open, honest discourse and rigorous empirical testing.