AI Glossary / Natural language and speech / Beginner

Article (part-of-speech)

Definition

An article is a grammatical word that helps indicate how a noun phrase refers to something in context. In English, the main articles are a, an, and the. A and an are indefinite articles, while the is definite. In NLP annotation schemes such as Universal Dependencies, articles belong to the determiner category rather than receiving a separate top-level part-of-speech tag.

What it means

Articles help a reader interpret whether an entity is being introduced or treated as identifiable. “A visitor arrived” can introduce a visitor; “the visitor waited” can refer to a visitor the reader can identify from context. These are useful patterns, not infallible new-versus-old rules.

This entry concerns the grammatical category, not a newspaper article or research paper. Languages differ in whether and how they use articles, so English article rules should not be treated as a universal template for translation or multilingual text analysis.

How it works

An article forms part of a noun phrase. It can precede intervening modifiers, as in “the unusually quiet room,” rather than appearing immediately beside the noun. A tagger identifies the article's grammatical role, while a parser can connect it to the noun phrase it helps determine.

Definiteness concerns identifiability in context, not a guarantee that the entity was explicitly mentioned earlier. Shared circumstances or a unique description may make a referent identifiable. Conversely, indefinite expressions can be used in several ways, including introducing a particular entity or describing a nonspecific possibility.

The Universal Dependencies convention tags English articles as DET and marks their article status with an additional feature. Other determiners, such as this or some, are not thereby articles. The broader category and its subtype should be distinguished when interpreting a model's labels.

Three practical examples

1. Writing assistance: A tool can flag a potentially missing article in ordinary English prose, while considering whether the text is a heading, label, name, or fragment where omission may be intentional.

2. Reference analysis: “A device failed. The device was replaced” gives useful clues for linking mentions, although article choice alone is insufficient to establish identity.

3. Translation: A system translating into English may need to supply an article even when the source language expresses reference differently. The appropriate choice depends on context, not a word-for-word substitution rule.

Think of it this way

Think of an article as a small instruction about how to find the intended item in the conversation: introduce one, or identify the relevant one. The analogy is limited because article use is more varied than a filing command. General statements, names, institutional expressions, and shared context can behave differently, and the article alone does not identify the actual entity.

Worked example

Consider this invented passage: “A courier delivered a parcel. The courier left the parcel at reception.” There are four article occurrences: two instances of a in the first sentence and two instances of the in the second.

Under the intended reading, the first sentence introduces a courier and a parcel. The second sentence refers to the same two entities. A simple annotation can therefore pair a courier with The courier and a parcel with the parcel. The four article tokens participate in four noun-phrase mentions, but the passage describes two relevant entities, not four.

Now change the second sentence to “The manager left the parcel at reception.” The manager may be identifiable from shared workplace context, even though no manager was mentioned in the first sentence. The definite article does not prove that manager corefers with courier.

This illustrates a useful separation: article detection can be straightforward, while discourse reference still requires interpretation. The counts and links are a declared teaching analysis, not predictions from a parser or evidence that every definite noun phrase has an earlier textual antecedent.

Illustrative case study

Fictional case: Cedar Editing develops a grammar assistant for internal documents. It applies ordinary sentence rules to every line and adds articles to product labels and table headings where the writers intended compact noun phrases.

The team distinguishes prose from interface labels and asks reviewers to assess edits in context. It also avoids replacing an indefinite article with a definite one merely to make wording sound more polished. The fictional case illustrates an editing-design issue and reports no measured improvement or universal rule for all writing styles.

Common misconceptions and mistakes

Do not assume the always points to an earlier noun phrase. An entity can be identifiable from shared knowledge or the situation. Also avoid assuming a always means the writer has no particular entity in mind; definiteness and specificity are related but different ideas.

Do not treat all determiners as articles. This, each, and many can be determiners under a tagging scheme without belonging to the article subtype. Finally, an article's short length does not make it semantically irrelevant: changing a to the can change what the reader is expected to identify.

Business use

Article handling matters in editing, translation, reference resolution, and language-learning products. It helps systems preserve how a writer introduces and tracks entities instead of treating noun phrases as isolated keywords.

Evaluate edits with their surrounding text and intended genre. A grammatically plausible insertion can still change the reference or sound unnatural in a heading. For multilingual systems, document the source and target languages' conventions and involve suitable language reviewers, since an English-trained article heuristic will not transfer uniformly to languages with different grammatical systems.

People also ask

Are a and an different in definiteness?

No. Both are indefinite articles in English; their distribution is conditioned by the following pronunciation rather than a definite-versus-indefinite distinction.

Is every determiner an article?

No. Articles form a subset of the broader determiner category used in many grammatical and NLP schemes.

Does a noun always need an article in English?

No. Proper names, many plural or mass-noun uses, and particular constructions can occur without one. Context and construction determine the appropriate form.

Check your understanding

A tagger labels the word the as DET. Does that contradict calling it an article?

Answer: No. In Universal Dependencies, DET is the broader part-of-speech category, and articles are a subtype. The two descriptions operate at different levels of classification.

Related terms

  • Determiner
  • Noun phrase
  • Anaphora
  • Part of speech
  • definiteness
  • reference
  • grammatical feature
  • specificity

Learn next

Compare articles with demonstratives and quantifiers, then study definiteness versus anaphoricity. Practice examples where the introduces an identifiable entity that has not been mentioned explicitly before.

More in Natural language and speech

  • Argumentative relations - Argumentative relations are the connections that show how one argument component bears on another, such as providing support or raising an objection.
  • Argumentative zoning - Argumentative zoning identifies the rhetorical role of passages, especially sentences, in scientific writing.
  • ARPAbet - ARPAbet is a system for representing speech sounds using ordinary ASCII letters and symbols, widely associated with English pronunciation resources.
  • Articulatory phonetics - Articulatory phonetics studies how speech sounds are produced by the vocal organs, including the lips, tongue, jaw, velum, and larynx.
  • Articulatory speech recognition - Definition: Articulatory speech recognition uses information about how speech is produced by the vocal tract to help identify spoken content.
  • Articulatory synthesis - Articulatory synthesis generates speech by modeling aspects of how the human vocal system produces sound.

Sources

Sources: Universal Dependencies, DET guidelines, read for article classification and the article feature: https://universaldependencies.org/u/pos/DET.html . Jurafsky and Martin, Chapter 24, read for definite and indefinite reference in discourse: https://web.stanford.edu/~jurafsky/slp3/24.pdf . Examples and counts are original; the entry focuses on English while distinguishing cross-language and annotation-scheme variation.

Previous: Array-sequenced organization Next: Articulated body pose estimation

  • Home
  • Daily AI Updates
  • Blog
  • AIFBL Academy
  • AI Glossary
  • Products and Tools
  • Store
  • Free CFO AI Prompt Pack
  • About Shihan Sheriff
  • Privacy Policy
  • Terms of Use
  • Refund Policy
  • Cookie Policy
  • Earnings Disclaimer