When the Times Crossword’s cryptic clue “Brittle‑hearted lover, initially (4)” appeared on a rainy Tuesday in March 2023, a 34‑year‑old software engineer in Manchester scribbled “ROMEO” in the margin, laughed, and moved on. A week later, the same engineer earned a 14‑point boost on the verbal‑analogy section of the UK’s national IQ assessment, outscoring his cohort by a full standard deviation. The coincidence sparked a small but growing body of research that treats cryptic crosswords not merely as pastime but as a cognitive training ground for the very skills that verbal analogies demand.
- Have you read these?
- What is a Spatial IQ test?
- The Evolution of IQ Testing: From Binet to AI-Driven Assessments
- What are the uses of iq test?
The Anatomy of a Verbal Analogy
Verbal analogy items—think “hand is to glove as foot is to ?”—require three intertwined operations: (1) rapid retrieval of lexical items, (2) mapping of relational structures, and (3) the ability to suppress irrelevant associations while homing in on the abstract link. In the WAIS‑IV, the “Similarities” subtest, which closely mirrors classic analogy formats, draws heavily on the left inferior frontal gyrus (LIFG) and the anterior temporal lobe (ATL). Neuro‑imaging work by Duncan and Owen (2000, University of Cambridge) showed that LIFG activation spikes when participants must select the correct relational term among competing alternatives. The same region lights up in studies of semantic fluency, a proxy for vocabulary depth (Bates et al., 2003, University of Edinburgh).
Cryptic Crosswords: A Mini‑Laboratory for Relational Reasoning
Unlike straight‑definition crosswords, cryptic clues embed a definition and a wordplay component—an anagram, a charade, a hidden word—within a single sentence. Solving them forces the brain to juggle divergent possibilities (the many ways a clue could be parsed) and convergent resolution (the single correct answer). This dual demand mirrors the cognitive choreography required for analogical reasoning.
Robert A. Björk and Elizabeth L. Björk highlighted the “desirable difficulty” of such puzzles in their seminal 2011 paper, “Making Things Hard on Yourself, but in a Good Way.” They argued that retrieval practice embedded in puzzles strengthens long‑term semantic networks, a claim supported by a 2017 experiment in Memory & Cognition where participants who completed daily crossword puzzles (n = 68) demonstrated a 9 % increase in vocabulary recall after four weeks compared with a control group (Bjork & Bjork, 2017, University of California, Irvine).
Crucially, the same study measured performance on a subset of the WAIS‑IV verbal analogies. The crossword group averaged 2.3 points higher (p
Neurocognitive Overlap: Evidence from Brain‑Imaging
A 2012 fMRI investigation by K. M. O’Connor at the University of Oxford mapped brain activity while participants solved cryptic clues versus standard definition clues. Both conditions activated the LIFG, but cryptic solving recruited additional bilateral posterior temporal regions associated with pattern‑completion and semantic integration (O’Connor et al., 2012, n = 24). The authors noted a 27 % increase in functional connectivity between LIFG and ATL during cryptic trials, a pattern that closely resembles the network observed in analogical reasoning tasks (Hampshire et al., 2012, Nature Neuroscience).
These findings dovetail with the “relational integration” model proposed by Greenberg, Barch, and Braver (2006, University of Michigan). Their work posits that the frontoparietal control network, especially the dorsolateral prefrontal cortex, orchestrates the alignment of relational schemas across domains. Cryptic crosswords, by demanding simultaneous interpretation of multiple linguistic constraints, appear to exercise this very circuitry.
Large‑Scale Behavioral Correlates
The Cambridge Brain Sciences (CBS) platform, launched by researchers at Cambridge University in 2014, offers a suite of online cognitive tasks to over 150,000 volunteers worldwide. In a 2020 analysis of CBS data (n = 12,487), participants who self‑identified as “regular cryptic crossword solvers” (≥3 puzzles per week) scored an average of 0.42 standard deviations higher on the “Word Analogies” task than non‑solvers (p
Similarly, a 2021 survey conducted by the UK Crossword Association (UKCA) collected responses from 4,312 hobbyists. Sixty‑eight percent reported feeling “confident” or “very confident” when faced with verbal‑analogy questions on standardized tests, compared with 42 % of a matched control group (UKCA, 2021). While self‑report data lack the rigor of experimental designs, the magnitude of the confidence gap aligns with the performance differentials observed in laboratory settings.
Why Vocabulary Matters
Vocabulary breadth is a well‑established predictor of performance on verbal analogies. The seminal meta‑analysis by Deary, Strand, Smith, and Fernandes (2007, University of Edinburgh) quantified a correlation of r = .71 between vocabulary scores and WAIS‑IV verbal comprehension. Cryptic crosswords provide a built‑in vocabulary expansion engine: anagrams and hidden‑word clues force solvers to retrieve less‑common lexical items, while the definition component reinforces semantic encoding. A 2015 longitudinal study by Dr. Susan A. Green (University of Toronto) tracked 120 university students over two semesters; those who added cryptic crossword practice (30 minutes, three times per week) increased their Peabody Picture Vocabulary Test scores by an average of 5.6 points, whereas the control group showed no significant change (Green, 2015).
From Lateral Thinking to Linear Logic
Cryptic crosswords also cultivate lateral‑thinking skills—seeing a clue from multiple angles before converging on a single solution. This habit of “mental set shifting” is measured by the Wisconsin Card Sorting Test (WCST), where higher switch‑cost efficiency predicts better analogy performance (Miyake et al., 2000, University of Colorado). In a 2018 experiment, participants who completed a week‑long cryptic‑crossword training protocol (n = 30) reduced their WCST perseverative errors by 18 % relative to a control group (Liu et al., 2018, Shanghai Jiao Tong University). The same participants improved their analogy scores by 3.1 points on a custom relational‑reasoning battery, reinforcing the causal chain from set‑shifting to analogical acuity.
Limitations and Open Questions
While the conver
