A Tiny Turkish Particle Reveals How Preschoolers Master Focus in Language

Ask a four-year-old whether every child in a picture is holding a balloon, and you may get a surprising answer. Even when every child in the image clutches a balloon, an extra balloon floating unheld in the corner can prompt an emphatic No. This stubborn error, known to linguists as quantifier spreading or q-spreading, has puzzled developmental scientists for decades. A new study of Turkish-speaking preschoolers now suggests a striking remedy hiding inside the grammar itself: a single movable question particle that, when attached to the right word, seems to redirect children’s attention and steer them toward the adult interpretation.

The research, conducted by Münir Özturhan and Utako Minai of the University of Kansas and published in Frontiers in Human Neuroscience, investigates whether young children can exploit morphologically marked focus during sentence comprehension. In Turkish, focus, the flagging of the most important information in a sentence, can be expressed not only through stress or word order but through the placement of the question particle mI. Typically, this particle sits at the end of a yes-no question. But when it attaches instead to a particular constituent, that constituent becomes the focus of the question, producing an interpretation equivalent to an English cleft sentence. A question meaning Is every child holding a balloon? can be transformed, by moving the particle onto the subject, into Is it every child that is holding a balloon?

This grammatical quirk makes Turkish an ideal laboratory for testing a long-standing debate in language acquisition. Researchers have split into two camps over children’s knowledge of focus. Partial competence accounts hold that young children genuinely lack the adult-like linguistic machinery needed to compute focus interpretations. Full competence accounts counter that children possess the underlying knowledge but struggle to deploy it, hampered by computational demands, task pressures, or attentional limitations. The evidence has been mixed. Children as young as toddlers modulate their intonation to mark focus in production, and longitudinal records show Turkish children using the question particle productively and accurately by around age two, including in the third-person singular configuration used in the study’s stimuli. Yet comprehension experiments, whether using offline judgments or online eye-tracking, have repeatedly found children failing to use focus cues that adults exploit effortlessly.

Previous work on focus particles such as English only illustrates the problem. When only appears with a subject, as in Only the cat is holding a balloon, children across many languages, including English, Japanese, Mandarin, German, European Portuguese, and Turkish itself, tend to misinterpret the sentence as if only modified the object or predicate. This subject-object asymmetry has been documented repeatedly, and no comparable study had ever tested whether Turkish children understand the focus function of the question particle in comprehension. Production data offered tantalizing hints: a three-year-old recorded in the CHILDES corpus responded to an experimenter’s prompt with Kediler mi?, roughly Is it the cats that scratch?, correctly attaching the particle to a subject to mark focus. Child-directed speech analyzed by Aylin Küntay and Dan Slobin shows Turkish mothers routinely using the particle in both positions, giving children abundant input. But production and input do not guarantee comprehension.

To sidestep this methodological gap, the researchers turned to universal quantification, a domain where children’s difficulties are well established and, crucially, where attention seems to be the bottleneck. When children hear a question containing her, the Turkish counterpart of every, paired with a picture showing each child holding a balloon plus one extra balloon held by nobody, many children reject the question. Their justification points to the orphan balloon, an object that is logically irrelevant to whether every child holds one. Under the full competence view of quantification, this error reflects a failure of attentional control rather than a deficit in semantic knowledge. One influential account, developed by Minai and colleagues, draws on the Attentional Inertia hypothesis: children lock onto the visually salient extra object and cannot shift perspective to evaluate the picture in terms of the characters and their possessions.

That account suggests a way to test focus. If visual salience can tug children’s attention toward the wrong element, perhaps linguistic focus marking can tug it back toward the right one. The researchers recruited 18 monolingual Turkish-speaking children aged four to nearly six, all attending Turkish-medium preschools in Turkey, along with 18 adult native speakers as controls. In a picture-sentence verification task, participants viewed extra-object pictures and answered pre-recorded yes-no questions. In the Baseline condition, the question particle appeared sentence-finally, yielding a broad-focus question like Is every child holding a balloon? In the critical Focused every condition, the particle attached to the quantified subject, producing Is it every child that is holding a balloon? The two conditions formed minimal pairs, differing only in particle position. Each child completed both conditions across two Zoom sessions spaced at least a week apart, with the Baseline always first. Children also had to justify each answer, allowing the experimenters to confirm that rejections were genuine q-spreading errors driven by the extra object.

The results were clear. Children answered adult-like on 68 percent of trials in the Focused every condition versus 58 percent in the Baseline condition, and adults performed near ceiling in both, confirming the stimuli were interpreted as intended. Because most children responded uniformly across conditions, nine were at ceiling and three at floor in both, the standard mixed-effects logistic regression produced unstable estimates, a statistical phenomenon known as quasi-complete separation. The team therefore turned to Bayesian modeling with weakly informative priors, which converged cleanly and confirmed the effect. The model estimated an average child would respond adult-like on 89 percent of trials in the focused condition versus 68 percent in the baseline, a difference of 21 percentage points, with a posterior probability of 0.997 that the difference was positive. A Bayes factor analysis strengthened the case further: the data were roughly 13.7 times more likely under a model including the focus manipulation than under one without it.

Individual patterns added an intriguing layer. Twelve children responded identically across both conditions, either always committing q-spreading or never doing so, replicating the bimodal splitting seen in earlier studies. Among the six children whose performance varied, four answered more accurately when the subject carried the focus particle, and not a single child did better in the baseline. Age did not reliably predict performance, though the small sample and narrow age range limit conclusions there. Filler items, which used mismatching pictures without extra objects, were answered at 94 to 99 percent accuracy, demonstrating that the children understood the task and could interpret universal quantifiers when the treacherous extra-object context was absent. The condition difference, in other words, was not a general comprehension failure but a targeted effect of where attention was directed.

The authors are careful about the limits of their claims. The study did not include a direct, standalone measure of focus comprehension, so the conclusion that children fully possess adult-like knowledge of the question particle as a focus marker rests on inference. The fixed session order leaves open a carryover explanation: perhaps the second-session improvement reflected general learning from prior exposure to the extra-object setup rather than the focus manipulation itself. The authors note, however, that previous q-spreading studies, including a within-subject eye-tracking study by Minai and colleagues and a longitudinal study by Aravind and colleagues, found little evidence that early exposure to extra-object contexts facilitates later performance; one longitudinal dataset even showed q-spreading increasing over time. Sample size and remote testing may also have suppressed overall accuracy, which at 0.68 in the focused condition was lower than in some comparable studies. A future condition focusing the object, which might cumulatively worsen q-spreading, remains to be tested.

Even with those caveats, the findings mark a first. Turkish-speaking preschoolers, previously shown only to produce the question particle as a focus marker, now demonstrably use it as a comprehension cue, employing a purely morphological device to guide attention toward the subject of a question and away from a salient but irrelevant distractor. The result bolsters the full competence view on two fronts at once, suggesting that children’s difficulties with both focus and quantification stem not from missing grammatical knowledge but from the challenge of allocating attention during real-time comprehension. It also expands the inventory of focus-marking devices known to be within a preschooler’s grasp, beyond the focus particles, prosody, and clefting studied in English, German, Japanese, and Mandarin. Sometimes, it turns out, the difference between a child’s error and an adult’s insight is a single two-letter particle, silently telling the young mind where to look.

Subject of Research: Morphologically marked focus and its role in preschool children's comprehension of universal quantifier questions in Turkish

Article Title: Examining children’s sensitivity to morphologically-marked focus in the comprehension of question sentences: evidence from quantifier comprehension in child Turkish

Article References: Examining children’s sensitivity to morphologically-marked focus in the comprehension of question sentences: evidence from quantifier comprehension in child Turkish. (n.d.). https://doi.org/10.3389/fnhum.2026.1914200

Image Credits: AI Generated

DOI: 10.3389/fnhum.2026.1914200

Keywords: language acquisition, focus marking, Turkish, universal quantifier, quantifier spreading, child language comprehension, morphological focus, question particle, psycholinguistics, developmental linguistics, attention, Frontiers in Human Neuroscience