Understanding the Bottleneck Hypothesis in Language Learning

This blogpost was first published in 2013, when I was exploring what makes some aspects of second language acquisition easier than others. At the time, I had just read Roumyana Slabakova’s chapter on the Bottleneck Hypothesis, which argues that functional morphology (tense, aspect, agreement markers and so on) is the hardest part of a new language to master. More than a decade later (I am writing this in 2025), I still find the question fascinating: Why do learners struggle with certain features of grammar while finding others more accessible? And what does this mean for language teaching methods such as communicative language teaching, task-based learning, or the more recent use of AI-powered practice tools?

I have kept the original text of the post intact, because it offers a snapshot of how I engaged with these questions in 2013. But I also add, further below, some updated reflections from 2025. These include more recent debates in second language acquisition, the role of plurilingual competence, and the impact of digital tools and artificial intelligence on what learners perceive as “easy” or “hard” to learn.

---

“What is easy and what is hard to acquire in a second language?”

What is easy and what is hard to learn in a foreign language? I just read a chapter in an edited volume¹ on this very question, and this post aims to provide both a synopsis of the chapter and a critical reaction to it.

What is the bottleneck hypothesis?

To answer the question, Slabakova (2013) draws on Universal Grammar theory (UG), to develop what she calls the Bottleneck Hypothesis. In brief, UG postulates that, of all the features that make up a language (e.g. word order, semantic relations, verb inflections etc.), some are hard-wired to the brain, and therefore shared across all human languages (i.e., they are linguistic universals), whereas others differ from language to language (or, to be technical, they are subject to parametric variation).

In her chapter, Slabakova presents readers with a robust set of empirical data, drawn from several experimental studies, and makes a compelling argument that what challenges learners most is functional morphology (roughly speaking, functional morphology refers to these parts of words which denote tense, person, aspect etc.). She argues that once functional morphology is mastered, it is easy to acquire other aspects of language, such as syntactic structures or semantic notions. Hence, functional morphology constitutes a ‘bottleneck’ of sorts in the process of second language acquisition.²

What are the implications of the bottleneck hypothesis for teaching?

Although UG is primarily concerned with understanding the cognitive processes of language acquisition rather than prescribing pedagogical practice, the chapter includes a discussion of implications for teaching (pp. 24-25). On the basis of her impressive empirical data, Slabakova argues that the communicative approach to language learning does not constitute an optimally effective way to learn a second language. She suggests that it is pedagogically beneficial to engage students in activities that highlight the hard-to-acquire morphological forms, and to provide ample opportunities for practice.

In her words:

In a sense, drilling of the functional morphology is inevitable if the form has to […] get sufficiently automatic for easy lexical access. […] Thus, the bottom line of the chapter is: Practice your functional morphology! In ample, clear unambiguous context! As in learning other lexical items, it may be painful, but – no pain, no gain! (p. 25)

A broader understanding of the communicative approach

It is my view that Slabakova is only partly correct in making this suggestion. Much of her criticism seems to focus on the earliest (or ‘strong’) formulations of the communicative approach, which tended to completely exclude the study of formal features of language.³ Notably, the article explicitly mentions seminal (though arguably dated) contributions, like Canale and Swain (1980) and Savignon (1983). This might suggest some disconnect from more recent thinking in the field of language education.

In pointing the deficiencies of these approaches, Slabakova makes a valid point, but it is a point that has already been addressed in language education. More recent pedagogical approaches (e.g., Willis & Willis 2007) focus on raising awareness of formal features, and do so in ways that align with current pedagogical thought in ways that the repetitive drilling activities that Slabakova seems to suggest don’t.

A broader agenda of language education

More importantly, I think Slabakova’s criticism of communicative theory brings to the forefront the question of ‘what features of language should we be learning?”. If the goal of second language pedagogy is to help learners to perfectly reproduce linguistic structures in the target language, then it is hard to argue against the kind of approach that Slabakova suggests.

Communicative language teaching, however, is not just about learning the rules through which sounds and words come together to produce meaning. Rather, it is about learning to use language (1) in context, (2) in order to do things. It aims, for instance, to help learners understand that “Could you pass me the salt, please?” is not really a question, and that “It’s stuffy in here” might be functionally equivalent to “Open the window!”

Deciding what language is appropriate to different contexts is highly culture-specific, and does not seem to flow easily from the mastery of morphological from. It is therefore hard to see how a pedagogy which is based primarily on mastering morphological variation might be helpful in developing skills necessary to use language effectively.

An appraisal of the bottleneck hypothesis

Coming back to the original question, the point I am trying to make is that Slabakova may well be right in identifying functional morphology as the bottleneck of second language acquisition. However, the cognitive challenges associated with mastering functional morphology do not necessarily promote it to the status of a learning priority. Learning priorities can only be determined on the basis of the learners’ specific needs, and this is a domain which UG is decidedly unsuited to inform.

---

Revisiting the bottleneck hypothesis in 2025

More than a decade has passed since I first wrote this post, and the debates around what is “easy” or “hard” in second language learning have become both more nuanced and more diverse. What follows, therefore, is not so much a revision of my earlier argument, but rather my attempt to re-position it alongside more recent discussions in applied linguistics and pedagogy.

Functional morphology revisited

Slabakova’s Bottleneck Hypothesis continues to provoke interest, but it is increasingly evident that the tidy picture offered by Universal Grammar cannot fully account for the complexity of learner experience. Insights from corpus linguistics and usage-based research, for example, show how frequency, salience, and input distribution shape acquisition. A learner’s difficulty with verb endings or tense-aspect markers is not only a matter of cognitive architecture but also of how often, and in what ways, these forms appear in authentic language use. Put differently, we cannot know with much certainty if the “bottleneck” is hardwired; it is just as likely that it is at least partially constructed by the linguistic environment and the affordances that emerge in it.

From drilling to alternative practices

Slabakova’s recommendation to “practice your functional morphology” went counter to pedagogical orthodoxy at the time, and has become even less tenable as the pedagogical landscape continued to shift over time. While repetitive drilling remains one possible strategy (and it may yield results for some people and some learning objectives), language teachers today often can draw on a wide range of alternatives that are less mechanical and more embedded in meaning.

Techniques such as input flooding, focus-on-form interventions, and structured noticing tasks allow learners to encounter difficult forms in ways that connect with communicative goals. More recently, intelligent digital tools, ranging from AI chatbots to automated feedback systems, create opportunities for personalised, targeted practice. These tools are not a panacea, but they offer possibilities beyond the “no pain, no gain” ethos of earlier drilling traditions.

Communicative competence and beyond

Finally, the discussion about what counts as a learning priority has widened considerably. Communicative Language Teaching is no longer just about accurate grammar in context, but about the ability to navigate meaning across languages and cultures. The Council of Europe’s CEFR Companion Volume has pushed us toward plurilingual and pluricultural competence: the capacity to draw on one’s full linguistic repertoire to make sense, to negotiate, and to act in diverse communicative settings. From this perspective, the bottleneck is not functional morphology at all, but the challenge of developing the awareness, flexibility, and confidence to use language as a social practice.

---

Concluding remarks

Looking back on my older writing, I am struck by how the question “what is easy and what is hard to learn in a second language?” resists a single answer. Slabakova’s identification of functional morphology as a cognitive bottleneck remains valuable, but the field has since taught us that difficulty is shaped as much by context, pedagogy, and social purpose as by internal grammar. What teachers and learners prioritise will always depend on the needs they bring to the classroom, the tools they have at their disposal, and the communicative worlds they inhabit. For me, then the bottleneck is not simply a universal feature of acquisition, but a shifting challenge that reflects the evolving realities of language education.

What about you, though? What do you find easy or hard in learning or teaching a new language? Do Slabakova’s ideas about functional morphology resonate with your experience, or do you see other “bottlenecks” in practice? Share your thoughts in the comments below —I’d love to continue the conversation.

---