tl;dr-ELT

Sometimes, research from developmental psychology feels oddly familiar. I was reading a new study exploring how infants map new words to unseen objects & I immediately thought of beginner English learners trying to guess what “burglar” or “corkscrew” might mean without a picture. How do we learn when we can’t see what’s being described?

That’s exactly what Luchkina & Waxman (2025) explored in their recent PLOS ONE paper. Their central question: Can infants learn the meaning of a novel noun even if they can’t see the object it refers to?

The Study

Two age groups were tested -12-month-olds & 15-month-olds (67 in each)- using an online eye-tracking platform called Lookit. The experiment involved:

• Priming infants with familiar objects from one semantic field (e.g. fruits: “apple, orange, banana”),
• Introducing a hidden object using a made-up noun (e.g., “Look, a modi!”) that infants couldn’t see,
• Testing whether infants could match this new word to a correct object later—choosing between two novel items, one from the same semantic neighbourhood (e.g., a dragon fruit) & one from a different category (e.g., an ottoman).

At 15 months, infants consistently looked longer at the object from the semantically related group- suggesting they’d inferred something about the meaning of “modi,” even though they’d never seen it before.

The 12-month-olds? Even with 7 days of vocabulary training using picture books, they didn’t succeed. Their eye movements suggested preferences for familiar categories (e.g. animals over tools), but not evidence of actual word learning in the absence of a referent.

Why Does This Matter?

This taps into something fundamental about language: its capacity to represent the unseen, the abstract, the not-here. This study is a small but powerful step toward understanding how humans come to learn about things they can’t see directly- whether that’s an object or a concept.

The study builds on earlier work showing that by 12 months, infants understand words for hidden but previously seen objects (e.g., Behne et al., 2012; Bohn et al., 2018), but it’s not until around 15 months that they can form representations for novel absent nouns.

It also refines findings from Ferguson et al. (2014), showing that semantic priming -rather than grammar alone -can support this leap. In a landmark study, Ferguson et al. showed that 19-month-old infants could use grammar- specifically, the animacy of a known verb—to make inferences about a novel noun. For example, if they heard “The dax is crying,” they inferred that “dax” must be an animate being, even though they couldn’t see it.

However, 15-month-olds in that study didn’t show this ability- suggesting a developmental limit or a lack of grammatical knowledge. The new study offers a refinement of Ferguson’s findings: it’s not that 15-month-olds can’t infer meanings in the absence of a referent- it’s that they may need the right kind of cue, & semantic context is more accessible than syntactic structure at this age.

Teacher Takeaways?

While this is early-stage, pre-linguistic research, it sparks a few reflections for the classroom:

• Semantic fields matter. Infants showed better learning when the novel word was introduced in a coherent semantic context. The same may apply to language learners of all ages. Grouping new vocabulary by topic (vs. random lists) might not just be easier—it may be how we’re wired to learn.
• Context supports abstraction. Learning a new word without seeing it? Hard—but not impossible if contextual cues help narrow the possibilities. For ELT, this reinforces the value of co-text, narrative & expectation-setting when introducing new, unfamiliar terms.
• Development takes time. Just as 12-month-olds aren’t ready to map new words to absent referents, our lower-level learners may not be ready for full-blown abstract vocabulary without visible support. Timing, readiness, & appropriate scaffolding matter.

How do you introduce unfamiliar vocabulary when the concept isn’t visible or tangible? Do you use metaphor, analogy, images—or just trust the context will carry it?