--- title: "Explorations in an English Poetry Corpus — paper by Jacobs (2018)" description: "Jacobs addresses a gap in empirical literary research. 10 references." authors: ["Arthur M. Jacobs"] year: 2018 venue: "arXiv:1801.02054" source: "https://arxiv.org/abs/1801.02054" doi: "10.48550/arxiv.1801.02054" arxiv: "1801.02054" licence: "arXiv — non-exclusive license to distribute" licence_url: "http://arxiv.org/licenses/nonexclusive-distrib/1.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/jacobs-2018-explorations-english-poetry-corpus-neurocognitive/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/jacobs-2018-explorations-english-poetry-corpus-neurocognitive/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2018](https://md.research.elamur.ai/free/en/papers/2018/page/6/) / Explorations in an English Poetry Corpus: A Neurocognitive Poetics Perspective # Explorations in an English Poetry Corpus: A Neurocognitive Poetics Perspective **Authors:** [Arthur M. Jacobs](https://md.research.elamur.ai/free/en/authors/arthur-m-jacobs/) · **Year:** 2018 · **Venue:** arXiv:1801.02054 · **Licence:** [arXiv — non-exclusive license to distribute](http://arxiv.org/licenses/nonexclusive-distrib/1.0/) **Links:** [Publisher's page](https://arxiv.org/abs/1801.02054) · [PDF](https://arxiv.org/pdf/1801.02054) · [DOI 10.48550/arxiv.1801.02054](https://doi.org/10.48550/arxiv.1801.02054) · [arXiv 1801.02054](https://arxiv.org/abs/1801.02054) · [OpenAlex](https://openalex.org/W2783804372) **Citations:** 0 citations (OpenAlex, 2026-10-14) ## Summary Jacobs addresses a gap in empirical literary research. Psycholinguistics and Neurocognitive Poetics lack large corpora built specifically from literary texts, so measures such as trigram surprisal are usually estimated from non-literary language. He assembles the Gutenberg Literary English Corpus (GLEC) from Project Gutenberg plus a Shakespeare collection. It holds about 3000 fiction and non-fiction works by 130+ authors. From it he carves out the Gutenberg English Poetry Corpus (GEPC). He cleaned it largely by hand, removing prefaces, footnotes, indices, prose passages and non-English poems, so that it can serve as a genuine poetry-only language model. He then shows Quantitative Narrative Analysis (QNA) on this subcorpus: - latent-semantic author similarity, displayed with multidimensional scaling; - NMF topic profiles per author; - a detailed comparison of George Eliot's "How Lisa Loved the King" and James Joyce's "Chamber Music" covering lexical diversity, keyness, collocations, sonority, valence and arousal. Pound and Joyce appear as outliers, and the Lake poets group together. The two compared poems turn out nearly identical in global sonority and affect. The resource matters as a training and test corpus and as a tool for building and controlling experimental stimuli. ## Key points - **GLEC scale:** roughly 3000 texts, amounting to some 250 million words spread over approximately 12 million sentences. It mixes over 900 novels, 500+ short stories, about 100 plays and 500+ non-fiction pieces, and apart from the poetry part it remains uncleaned. - **GEPC scale:** 116 texts merged into 47 per-author compound texts, totalling 1,808,160 tokens and 41,857 types. Lengths follow an exponential distribution with a median of about 23,000 words, from Byron at about 210k down to Alcott at under 400. - **Similarity pipeline:** a document-term matrix built with sklearn CountVectorizer (max document frequency .95), NLTK stopwords and SnowballStemmer, analysed with LSA and plotted with MDS. A separate NMF model extracted 20 topics. - **Topic breadth varies widely:** Whitman's poems touch only 4 of the 20 topics, while Dickens's verse spans 15. Tennyson's dominant topic is Arthurian (king, knight, lancelot, queen). - **Eliot vs Joyce:** | Measure | Eliot | Joyce | |---|---|---| | Tokens | 2702 | 1221 | | Type-token ratio | .50 | .53 | | Mean sonority | 5.19 | 5.26 | Valence and arousal are almost equal, and love is the central motif in both poems. - **Limitations and outlook:** the texts date from 1623 to 1952, with a median year of 1885 for the GEPC. Even so, the GLEC as a language model covered all 464 metaphors in an earlier surprisal study, some of them contemporary. ## Interesting ideas - **Genre-pure training slices:** the author cuts all prose (prefaces, footnotes, notes) out of the poetry texts because even small prose remnants skew n-gram statistics for the target register; register purity beats raw size when the model must represent one specific genre. - **Coverage ("hit rate") as a sanity check:** before trusting surprisal values, the author checks that the corpus model actually contains every test item, a cheap test of whether a resource fits a new domain. - **Topic breadth as a document profile:** counting how many topics a text touches (4 for Whitman vs 15 for Dickens) separates thematically narrow texts from sprawling ones with a single interpretable number. - **Lexical dispersion of a motif:** plotting where a key word appears across the whole text shows whether it is a constant thread or a local episode, a cheap structural cue to the central theme. - **Two routes to lexical affect:** valence and arousal can come either from human rating lists or from a word's association with prototypical emotion labels (Turney & Littman), and both yield per-word explainable scores. ## For Mova How these ideas could serve Mova, our small open language model — its files, stages and tests: - **I1** (learn): keep separate, cleaned register slices (fables, school prose, word problems) for silver data and evaluation, so statistics from one genre do not leak into another. - **I2** (learn): report, alongside accuracy, the share of test lemmas, verb frames and constructions that the training material actually covers, giving a natural control for out-of-domain failures. - **I3** (world): count how many distinct ontology domains a text's entities and events fall into, and flag sprawling texts where the world model may need more careful scoping. - **I4** (answer): when answering "what is this fable about", prefer entities or concepts whose mentions are spread across the text rather than bunched in one passage, and cite those positions as the explanation. - **I5** (frames): attach per-word valence to frame participants and outcomes, so that a mismatch between lexical tone and the event result can serve as an explainable irony cue. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* 📚 Shelves (7): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 9,086 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 8,889 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 9,296 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 9,206 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 744 works · [🤝 General Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/general-social-sciences/) · 87 works · [📚 Literature and Literary Theory](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/literature-and-literary-theory/) · 340 works 🧭 Topics (2): [English](https://md.research.elamur.ai/free/en/topics/english/) · 1,373 works · [🧒 Psycholinguistics & language acquisition](https://md.research.elamur.ai/free/en/topics/psycholinguistics/) · 66 works ## Cite Cite as: Arthur M. Jacobs. 2018. Explorations in an English Poetry Corpus: A Neurocognitive Poetics Perspective. arXiv:1801.02054. The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @misc{jacobs2018explorations, title = {{Explorations in an English Poetry Corpus: A Neurocognitive Poetics Perspective}}, author = {Arthur M. Jacobs}, year = {2018}, doi = {10.48550/arxiv.1801.02054}, eprint = {1801.02054}, archivePrefix = {arXiv}, url = {https://doi.org/10.48550/arxiv.1801.02054} } ``` - [← previous ECNU at SemEval-2018 Task 3: Exploration on Irony Detection from Tweets via Machine Learning and Deep Learning Methods](https://md.research.elamur.ai/free/en/papers/yin-2018-ecnu-semeval-2018-task-3/) - [next → Focalization in the Old Testament Narratives with Specific Examples from the Book of Ruth](https://md.research.elamur.ai/free/en/papers/nazarov-2021-focalization-old-testament-narratives-specific/) - [↑ back to papers of 2018](https://md.research.elamur.ai/free/en/papers/2018/page/6/) ## References 10 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Alexander Pope The Rape of the Lock and Other Poems, 1875 Is Mars Habitable?, 1907 [find](https://scholar.google.com/scholar?q=Alexander+Pope+The+Rape+of+the+Lock+and+Other+Poems%2C+1875+Is+Mars+Habitable%3F%2C+1907) 2. Ambrose Bierce A Cynic Looks at Life, 1912 [find](https://scholar.google.com/scholar?q=Ambrose+Bierce+A+Cynic+Looks+at+Life%2C+1912) 3. Andrew Lang Historical Mysteries, 1904 [find](https://scholar.google.com/scholar?q=Andrew+Lang+Historical+Mysteries%2C+1904) 4. Beatrix Potter The Eustace Diamonds, 1871 Turkey / A Past and a Future, 1917 The Tangled Skein, 1907 A Collection of Beatrix Potter Stories, 1902 Vivian Grey, 1826 [find](https://scholar.google.com/scholar?q=Beatrix+Potter+The+Eustace+Diamonds%2C+1871+Turkey+%2F+A+Past+and+a+Future%2C+1917+The+Tangled+Skein%2C+1907+A+Collection+of+Beatrix+Potter+Stories%2C+1902+Vivian+Grey%2C+1826) 5. Bret Harte Autobiography of Benjamin Franklin, Version 4, 1791 The Analysis of Mind, 1921 Dracula, 1897 The Queen of the Pirate Isle, 1886 The Expression of Emotion in Man and Animals, 1859 Oliver Twist, 1837 8 6 58 [find](https://scholar.google.com/scholar?q=Bret+Harte+Autobiography+of+Benjamin+Franklin%2C+Version+4%2C+1791+The+Analysis+of+Mind%2C+1921+Dracula%2C+1897+The+Queen+of+the+Pirate+Isle%2C+1886+The+Expression+of+Emotion+in+Man+and+Animals%2C+1859+Oliver+Twi) 6. DH Lawrence True Words for Brave Men, 1884 4 19 Jane Eyre, 1847 Women in Love, 1920 [find](https://scholar.google.com/scholar?q=DH+Lawrence+True+Words+for+Brave+Men%2C+1884+4+19+Jane+Eyre%2C+1847+Women+in+Love%2C+1920) 7. Edgar Allen Poe The Masque of the Red Death, 1842 Tarzan of the Apes, 1912 [find](https://scholar.google.com/scholar?q=Edgar+Allen+Poe+The+Masque+of+the+Red+Death%2C+1842+Tarzan+of+the+Apes%2C+1912) 8. Edmund Burke Burke's Speech on Conciliation with America, 1775 The Zeppelin's Passenger, 1918 [find](https://scholar.google.com/scholar?q=Edmund+Burke+Burke%27s+Speech+on+Conciliation+with+America%2C+1775+The+Zeppelin%27s+Passenger%2C+1918) 9. Elizabeth B Browning Sonnets from the Portuguese, 1850 [find](https://scholar.google.com/scholar?q=Elizabeth+B+Browning+Sonnets+from+the+Portuguese%2C+1850) 10. Emily Bronte Wuthering Heights, 1847 [find](https://scholar.google.com/scholar?q=Emily+Bronte+Wuthering+Heights%2C+1847) *Updated 24 October 2026*