Dictionaries vs. Glosses: The Fastest Way to Look Up a Word While Reading

By Beeblio · November 13, 2026

You are three paragraphs into a French short story. The sentence is elegant, the plot is pulling you forward — and then you hit an unknown word. What do you do next matters more than almost any other micro-decision in language learning. Do you tap the word for an instant gloss, or do you open a dictionary app, type or paste the word, scroll past a cascade of definitions, pick the right one, then find your place again? The two paths feel almost identical in effort. The research says they are not — not for comprehension, not for retention, and not for the long-term growth of your vocabulary.

The Coverage Problem: Why Every Unknown Word Counts

Before comparing glosses and dictionaries, it helps to understand why looking anything up matters so much. Vocabulary researchers have spent decades measuring how much of a text a reader must know in order to understand it. Studies of lexical coverage and comprehension have found that, for reading, 95% and 98% are the minimal and optimal coverages, respectively. In practice, successful reading generally requires 95–98% word coverage, with 95% allowing adequate comprehension with support — such as dictionaries or glosses — and 98% supporting unassisted understanding.

Those percentages translate into real vocabulary sizes. If 98% coverage of a text is needed for unassisted comprehension, then an 8,000- to 9,000-word-family vocabulary is needed for comprehension of written text. At the 1,500-word band — a common milestone for motivated beginners and early intermediates in a Romance language — most general texts will put you somewhere in the 94–97% coverage range, depending on the register. That means every lookup tool in your toolkit earns its keep. The question is which one earns it best.

To make this concrete, here is a short French passage written at the 1,500-word frequency band, with a small number of words just beyond it:

At the 1,500-word band, a learner reading that passage already recognises 96% of its tokens — roughly one unknown word per two lines. That is comfortably in the zone where comprehension support (a gloss, a dictionary, context inference) can bridge the gap between confusion and understanding. The practical question becomes: which form of support does it fastest and sticks longest?

What Is a Gloss, Exactly?

In second-language acquisition research, a gloss is a brief meaning aid attached directly to a target word — typically a short L1 translation or a simple L2 definition placed in the margin, as a pop-up, or inline in a digital text. Glosses are not new: marginal annotations appear in medieval manuscripts. What is new is the evidence base about how they work, and digital delivery that makes them frictionless.

The location of the gloss makes it easier to use than the dictionary, while supplying definitions or explanations in the text being read is believed to minimise students looking up word meanings in a dictionary, thus reducing reading interruptions. That reduction in interruption is not trivial. If a dictionary imposes excessive cognitive load on its users during the search process, it may reduce learners' attention to the target words, disrupt fluent reading, and discourage learners from searching for unfamiliar words.

Glosses sidestep that problem by design: the meaning arrives at the word, rather than the reader leaving the word to find the meaning.

What the Research Shows About Comprehension

The comprehension evidence for reading with glosses is broadly positive, though not perfectly uniform. Results of quantitative analyses have indicated that the second-language gloss condition significantly affected students' reading comprehension, while think-aloud protocols revealed that both types of glossing made reading comprehension smoother and faster than was possible for those who read without glosses.

Research has also found that collocational glossing — providing meaning for multi-word units rather than single words — promoted reading comprehension significantly compared to other reading conditions. This matters because language is not just a string of individual words: a learner who knows prendre and garde separately may still be mystified by prendre garde. A well-designed gloss can flag the collocation as the unit of meaning.

The picture is admittedly not monolithic. Some studies indicate that glossing significantly facilitated learning of target constructions while having no statistically significant influence on reading comprehension scores — often because control groups also performed well, suggesting a ceiling effect when texts are already close to the learner's level. This is worth knowing: glosses are not magic. If your text is already at 98%+ coverage, the marginal comprehension benefit of a gloss shrinks. Where glosses most reliably help is in the 94–97% range — exactly where a 1,500-word French learner typically lands with authentic material.

What the Research Shows About Retention

Comprehension during reading is only part of the story. The goal most learners actually care about is remembering the word tomorrow, next week, next month. Here the evidence for glosses is striking.

A meta-analysis in Studies in Second Language Acquisition investigated the overall effects of glossing on L2 vocabulary learning from reading across 42 studies (N = 3,802). Glossed reading led to significantly greater learning of words — 45.3% and 33.4% on immediate and delayed posttests, respectively — than non-glossed reading (26.6% and 19.8%). That is a large and durable gap. Even on a delayed test, glossed readers retained words at nearly double the rate of unglossed readers.

The seminal field study here is Hulstijn, Hollander, and Greidanus (1996), published in The Modern Language Journal. Dutch advanced students of French read a French short story under one of three conditions: marginal glosses, dictionary use, or control (no support). After reading, students were tested for their recall of 16 words that had appeared either once or three times in the text. The study found support for the hypothesis that frequency of occurrence fosters incidental vocabulary learning more when readers are given meanings through marginal glosses or when they look up meanings in a dictionary than when no external information is available. Critically, that study found that marginal glosses significantly improved retention compared to paper dictionaries.

Among gloss formats, multiple-choice glosses proved the most effective, while in-text glosses and glossaries were the least effective. L1 glosses — a translation into the learner's native language — yielded greater learning overall than L2 glosses. This last finding has practical implications: for a French learner whose native language is English, a gloss that shows the English equivalent is on average more effective than one that offers a French definition — at least for lower-to-intermediate proficiency levels.

The Dictionary's Real Cost: Cognitive Load

None of this means dictionaries are useless. Glosses better promote incidental vocabulary learning partly because in some cases students simply do not use the dictionary when given the choice. When dictionaries are indeed used, the results can be as good as with glosses. The problem is the "when" — learners frequently skip the lookup entirely because the friction is too high.

The use of a dictionary has been shown to have a positive effect on incidental vocabulary learning, yet studies show that L2 readers often decide not to use the dictionary when meeting unfamiliar words in a text. One reason reported by students is the time involved in flicking through dictionary pages and the subsequent disruption of the flow of reading.

Even with an instant electronic dictionary, the problem does not fully disappear. With the help of technology, dictionary users can now instantly access a meaning list using a mouse click. However, they must still spend great cognitive effort identifying the most appropriate meaning for a given context, contributing to the disruption of the flow of reading and decreasing the positive effect of dictionary use.

Consider what a typical French dictionary entry for a mid-frequency word looks like: four or five definitions, numbered sub-senses, example sentences, usage notes. A fluent reader uses all of that richness to disambiguate. An intermediate learner stalls on it. The cognitive tax is real, and it compounds across a reading session. Looking through irrelevant meanings not only increases extraneous cognitive load and impedes the flow of reading, but also discourages users from searching for unknown words, which in turn may negatively affect text comprehension.

The Format Details That Actually Matter

Not all glosses are created equal. The research points to several design factors that determine whether a gloss actually works.

Gloss type: marginal, pop-up, or multiple-choice?

Pop-up (on-demand) glosses have an important advantage over always-visible marginal glosses: they force the learner to make a choice to look. That minimal decision — do I know this word? — is a metacognitive act that strengthens encoding. The meta-analysis confirms that glossed reading led to significantly greater learning, and among formats, multiple-choice glosses — where the learner selects from several possible meanings — were the most effective, while in-text glosses and glossaries were the least effective. The requirement to discriminate between options appears to deepen processing.

L1 vs. L2 gloss language

Research indicates that L1 glosses generate deeper comprehension and recall ability in intermediate learners than L2 glosses, while inversely, advanced learners perform better when given L2 glosses. This suggests a staged approach: use L1 translations at lower bands, switch to L2 definitions as your vocabulary grows. At the 1,500-word level for French, L1 glosses (into English, or whichever your native language is) are typically the more efficient choice.

Electronic vs. paper

Research on computer-assisted language learning (CALL) finds that both L1 and L2 electronic glossing can be beneficial to learners' vocabulary development, and that electronic glosses lead to more learning gains than paper-based glosses. The speed of access matters: a 200-millisecond tap-to-gloss interaction has a very different cognitive cost than a 15-second paper-dictionary lookup.

The Role of Repetition and Spaced Review

Both glosses and dictionaries share one fundamental limitation: a single encounter with a word rarely produces lasting memory. The Hulstijn et al. research confirmed that frequency of occurrence fosters incidental vocabulary learning — when a word reappears, the form-meaning connection in the reader's mental lexicon is reinforced. But authentic texts do not reliably re-expose you to the same words at the right intervals.

This is where spaced repetition enters the picture. German psychologist Hermann Ebbinghaus investigated the "forgetfulness curve," a graphical representation of the exponential decline in information retention as a function of elapsed time — and his work has been replicated as recently as 2015 with similar results. His research showed that by revisiting new information at specific time points along the forgetfulness curve, learners can blunt its steep decline.

The practical payoff is substantial. A review of 317 experiments found higher recall after spaced practice than after massed practice: 47.3% against 36.7% on average. For vocabulary specifically, research shows that vocabulary learned through spaced repetition is retained more successfully than vocabulary learned through rote memorisation, and that regular repetition helps transfer information from short-term to long-term memory.

The implication is clear: the moment of a gloss lookup is the beginning of learning, not the end of it. For the word to stick, it needs a scheduled return.

Putting It Together: A Practical Framework

The research converges on a hierarchy of lookup tools, ranked roughly by their effect on both comprehension and retention:

1. On-demand pop-up gloss (L1 translation, for intermediates and below) — lowest disruption, good retention, especially when multiple-choice format forces active choice.
2. On-demand pop-up gloss (L2 definition, for upper-intermediates and above) — slightly higher cognitive effort, which can deepen encoding; appropriate once your French is strong enough to process a French definition reliably.
3. Electronic dictionary (monolingual or bilingual) — higher cognitive load, but rich information useful when nuance matters or when the gloss is insufficient. Research suggests monolingual dictionary use is more effective than bilingual dictionary use for vocabulary learning, though for lower bands, the cognitive cost of a monolingual L2 dictionary often outweighs its benefit.
4. Paper dictionary — maximum disruption, minimum usage rate. Worth keeping for dedicated study sessions, not for in-flow reading.

The table below summarises the key trade-offs:

What the Research Does Not Fully Settle

Honest reporting requires noting the limits. The glossing literature has several recurring methodological constraints: most studies are short (a single reading session), use captive student populations, and measure immediate or short-term retention rather than long-term acquisition. The finding that glossed readers outperform non-glossed readers on delayed tests is robust, but most delayed tests happen one or two weeks out — not six months out, which is the timescale that actually matters for a self-taught learner.

There is also a real tension between comprehension and acquisition. Research shows a possible trade-off effect of glosses: they help comprehension, but that ease may reduce the deeper processing — such as inference from context — that consolidates long-term vocabulary knowledge. Struggling to work out a word's meaning from context, while costly in the moment, may sometimes produce a stronger memory trace than receiving the answer instantly. This is not an argument against glosses; it is an argument for mixing strategies — using glosses to protect reading flow on most unknowns, and occasionally forcing yourself to infer before tapping.

FAQ

Should I always use a gloss instead of a dictionary?

Not always. Glosses win on speed and flow-preservation, and the meta-analytic evidence favours them for retention on delayed tests. But a dictionary is the better tool when you need to understand a word's full range of meanings, its grammar, or its collocational patterns — information that a gloss deliberately strips away for brevity. Use glosses for in-flow reading; use the dictionary for deliberate post-reading study of high-value words.

Does it matter whether the gloss is in my native language or in French?

Yes, and your proficiency level should guide the choice. L1 glosses generate deeper comprehension and recall in intermediate learners, while advanced learners tend to perform better with L2 glosses. As a rough rule: if you are at the 1,500-word band or below, prefer L1 (native-language) glosses. As you move toward 3,000–4,000 words, begin experimenting with L2 definitions.

What percentage coverage do I need before reading is enjoyable?

Minimal text comprehension requires at least 95% known-word coverage, while independent comprehension requires approximately 98%. Enjoyment typically tracks comprehension: at 96% — the level our sample French passage demonstrates for a 1,500-word learner — you will follow the gist clearly, encounter a few unknown words per paragraph, and find reading effortful but rewarding rather than exhausting.

How many times do I need to see a word before I remember it?

The number varies by word and by learner, but the vocabulary acquisition literature commonly cites roughly 10–15 meaningful exposures for a word to reach solid receptive knowledge — with spaced encounters being far more effective than massed ones. A single glossed encounter is a first exposure, not a learned word. That is why scheduling review matters as much as the initial lookup.

Are electronic glosses better than paper ones?

Research finds that electronic (CALL) glosses lead to more learning gains than paper-based glosses. Speed of access is likely a key mechanism: the faster a gloss appears, the less the reader's attention drifts from the sentence, and the more tightly the form-meaning connection is forged in context. Electronic glosses can also support multimedia — audio, example sentences, images — which can further strengthen encoding for some word types.

Sources

• Hulstijn, J. H., Hollander, M., & Greidanus, T. (1996). Incidental vocabulary learning by advanced foreign-language students: The influence of marginal glosses, dictionary use, and reoccurrence of unknown words. The Modern Language Journal, 80(3), 327–339.
• Yanagisawa, A., Webb, S., & Uchihara, T. (2020). How do different forms of glossing contribute to L2 vocabulary learning from reading? A meta-regression analysis. Studies in Second Language Acquisition, 42(2), 411–438.
• Nation, I. S. P. (2006). How large a vocabulary is needed for reading and listening? Canadian Modern Language Review, 63(1), 59–82.
• Schmitt, N., Jiang, X., & Grabe, W. (2011). The percentage of words known in a text and reading comprehension. The Modern Language Journal, 95(1), 26–43.
• Ko, M. H. (2005). Glosses, comprehension, and strategy use. Reading in a Foreign Language, 17(2), 125–143.
• Dang, T-D., Chen, G-D., et al. (2013). RoLo: A dictionary interface that minimizes extraneous cognitive load of lookup. Computers & Education, 61, 251–260.
• Kim, Y., et al. (2023). Effects of announcing a vocabulary test before reading a glossed text on reading behaviors and vocabulary acquisition: An eye-tracking study. Frontiers in Psychology.
• Laufer, B. (2014). Lexical coverage. In The Encyclopedia of Applied Linguistics. Wiley-Blackwell.
• Cepeda, N. J., et al. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354–380. (Summarised at vocabularybuilding.org.)

In Beeblio, every text you read is pinned to your measured frequency band — so if you are at 1,500 French words, you are always reading at the 96% coverage zone shown in the sample above, with a controlled number of new words just beyond your current range. When you tap an unknown word, you receive an instant contextual gloss, and the word is automatically queued in your spaced-repetition deck — turning the single moment of lookup into the first scheduled step of real retention. The research on reading with glosses suggests that this combination of level-appropriate text, frictionless in-context glossing, and spaced review is about as close as we currently get to an evidence-based reading workflow.