“Corpora” is simply the plural of “corpus,” a Latin word meaning “body.” It shows up in two very different fields: linguistics, where a corpus is a large collection of texts used to study language, and anatomy, where various body structures carry names like corpora cavernosa or corpora quadrigemina. Which meaning applies depends entirely on context, but both trace back to that same Latin root.
Corpora in Linguistics
In modern linguistics, a corpus is a collection of real-world texts, written or spoken, stored in digital form so researchers can search and analyze them. The plural, corpora, refers to multiple such collections. A corpus isn’t just any pile of text. It has four widely accepted qualities: the texts are machine-readable, they come from authentic (naturally occurring) language rather than made-up examples, they’re sampled systematically, and they’re chosen to be representative of a particular language or dialect.
The scale of these collections can be enormous. The Corpus of Contemporary American English (COCA), one of the largest freely available English corpora, contains roughly 450 million words drawn from spoken language, fiction, magazines, newspapers, and academic writing. The Corpus of Global Web-Based English (GloWbE) is even larger at 1.9 billion words pulled from 1.8 million web pages across 20 English-speaking countries. But corpora don’t have to be massive. Even a small collection of 10,000 words can yield useful patterns about how people actually use language.
Types of Corpora
Different corpora serve different purposes. A general corpus like COCA aims to represent a broad cross-section of a language. A historical corpus like the Corpus of Historical American English (COHA), with its 400 million words, lets researchers track how English has changed over decades or centuries. Learner corpora collect writing and speech from people studying a second language, giving teachers and researchers concrete evidence of common errors and how students improve over time. Parallel corpora contain the same texts translated into two or more languages, which is invaluable for translation research.
How Corpora Are Built
Building a corpus starts with deciding what language variety it should represent, then sampling texts accordingly. Written material is relatively straightforward to collect since most writing already exists in digital form. Spoken language is harder. Conversations and speeches need to be recorded and transcribed, though voice recognition software has made this faster for certain types of speech like monologues. Once the raw text is gathered, researchers often add annotation: tags that mark parts of speech, sentence boundaries, speaker overlaps in conversation, or other linguistic features that make the data searchable in meaningful ways.
Corpora and AI Training
Large language models, the technology behind modern AI chatbots and text generators, depend on enormous text corpora for training. These models learn to predict the next word in a sequence by processing vast amounts of written language. In a sense, an AI model trained on billions of words has “seen” more language than any individual human ever could. That massive exposure to real text is what allows these systems to produce coherent, natural-sounding output.
Corpora in Human Anatomy
In medicine and biology, “corpora” appears in the names of several distinct structures throughout the body. Each one refers to a “body” of tissue with a specific form and function.
Corpora Cavernosa
The corpora cavernosa are two columns of spongy erectile tissue that run side by side through the shaft of the penis, forming most of its bulk. They’re separated in the center by a thin wall of tissue called the intercavernous septum and enclosed by a thick fibrous outer layer. The interior has a Swiss-cheese appearance, full of interconnected spaces (sinusoids) woven between smooth muscle and connective tissue. During arousal, blood vessels within the corpora cavernosa fill these spaces with blood, producing an erection. A third, smaller cylinder called the corpus spongiosum (singular, since there’s only one) surrounds the urethra and sits beneath the paired corpora cavernosa.
Corpora Quadrigemina
Located on the back surface of the midbrain, the corpora quadrigemina are four small rounded bumps arranged in two pairs. The upper pair, called the superior colliculi, process visual information. They sit just below the thalamus, near the pineal gland. The lower pair, the inferior colliculi, are slightly smaller and handle auditory information. Together, these four structures act as relay stations, helping the brain coordinate reflexive responses to things you see and hear, like turning your head toward a sudden noise.
Corpora Arenacea
Sometimes called “brain sand,” corpora arenacea are tiny calcium deposits that form in the pineal gland, the small structure deep in the brain that produces melatonin. These mineral deposits appear early in life and tend to increase in number up through age 30 and beyond. They show up clearly on brain scans and are generally considered a normal part of aging rather than a sign of disease. In older individuals, larger deposits may develop a layered, laminated appearance.
Corpora Amylacea
Corpora amylacea are small, round, starch-like clumps that accumulate in the brain over time. They range from about 10 to 50 micrometers in diameter (roughly the width of a fine human hair) and are made mostly of sugars, with a small protein component and a calcium-based core. Small numbers appear in the brains of healthy aging individuals, but they become far more abundant in people with neurodegenerative conditions like Alzheimer’s disease, Parkinson’s disease, and ALS. In Alzheimer’s patients, they tend to concentrate in the entorhinal cortex and hippocampus, brain regions critical for memory. They’ve also been found in large quantities in some patients with temporal lobe epilepsy, where dense deposits can replace normal brain tissue in certain areas. Their exact role in disease progression remains unclear, but their presence is closely linked to both aging and neurodegeneration.
Corpus vs. Corpora: Getting the Grammar Right
Because “corpus” comes from Latin, its plural follows Latin rules: one corpus, two or more corpora. You’ll occasionally see “corpuses” used as an anglicized plural, and most dictionaries accept it, but “corpora” is standard in both academic and medical writing. If you searched “what is a corpora,” you were likely encountering the plural form and wondering what it meant. In everyday use, you’re most likely to run into it in linguistics courses, medical textbooks, or discussions about AI training data.