How lexicographers distinguish and document homonyms

A homonym is a word form that coincides with another form but carries a different meaning. The overlap may occur in spelling, pronunciation, or both. English examples include bank, referring to a financial institution or the side of a river, and bark, referring to tree covering or a dog’s sound. For a lexicographer, however, recognising a pair is only the beginning. The real work is deciding whether the meanings belong in one dictionary entry or in separate entries, and then presenting that decision clearly.

This judgement affects dictionaries, translation, language teaching and digital search tools. It is especially important in languages with rich systems of inflection, extensive borrowing and limited published reference material. A professional profile such as T.J. Otlogetswe’s language work shows how lexicography, corpus linguistics, translation and language documentation overlap. The method must respect actual usage while also making the structure of the vocabulary understandable to readers.

Establishing what kind of overlap exists

The first task is to classify the relationship between the forms. Homographs are written alike, while homophones sound alike; some words are both. Lead is a familiar English example because its pronunciation changes according to meaning. Flour and flower are homophones, but their spellings and histories remain distinct. A dictionary may cross-reference them, yet it should not imply that they share a single semantic origin.

Lexicographers also separate homonymy from polysemy. Polysemy occurs when related senses develop from a common meaning, as with head for a body part, a leader, or the top of something. Homonymy is more likely when meanings have separate histories or grammatical behaviour. The boundary is not always absolute, and historical evidence may be incomplete. A practical dictionary therefore considers etymology alongside present-day usage rather than allowing etymology to dictate every editorial decision.

Pronunciation and grammatical category provide useful tests. If one form is a noun and another is a verb, that difference may support separate entries, particularly when the meanings are unrelated. Inflection can reveal further distinctions: two identical-looking forms may belong to different paradigms or behave differently in a sentence. In Setswana and other languages, agreement markers, noun classes and verbal extensions can offer evidence that a surface resemblance hides separate lexical items.

Building evidence from real language

No responsible entry should rest on intuition alone. A lexicographer gathers examples from spoken language, newspapers, books, broadcasts, social media and specialist publications. A corpus reveals which meanings are common, which grammatical patterns accompany them and which collocations help readers identify the intended sense. The surrounding words often resolve an ambiguity that an isolated headword cannot.

For Australian English, a corpus might show how bark appears in discussions of eucalyptus trees, dogs, or Indigenous plant knowledge. It can also capture local vocabulary such as arvo, bush and mate, whose meanings depend partly on register and social setting. Evidence from Sydney news, regional Queensland speech and conversations in Perth may expose differences that a small, metropolitan-only dataset would miss. The aim is not to declare one variety correct, but to describe usage accurately and label it where necessary.

Fieldwork remains essential when a language has limited digitised material. Recordings, elicitation sessions and community-authored texts can supply examples that a general corpus lacks. A researcher must record who produced each form, in what setting, and with what intended meaning. Speakers may also explain that two forms considered identical by outsiders differ in tone, vowel length, stress or context. Such observations can prevent a dictionary from creating a false homonym.

Tracing history without forcing the answer

Etymology helps determine whether identical forms arrived through separate routes. A word may have been inherited from an earlier stage of the language, borrowed from a neighbouring language, or introduced through religious, commercial or administrative contact. In multilingual regions, two unrelated words can converge in pronunciation after generations of adaptation. Historical dictionaries, older translations and comparative evidence can make that process visible.

The history of an expression can also clarify why a modern sense seems surprising. Research into an ancient drinking saying, for example, demonstrates how a phrase can move across languages and centuries while retaining or reshaping its associations. A lexicographer treats such material as evidence, not as an entertaining story added after the definition. The entry should state what is well supported and avoid presenting speculation as fact.

Australian English offers many cases where contact and borrowing deserve careful treatment. Place names and vocabulary connected with Aboriginal and Torres Strait Islander languages may have multiple spellings, local pronunciations and contested explanations. A dictionary should consult authoritative community sources and acknowledge regional variation. It should never assume that a familiar Australian pronunciation gives the full history of a word.

Designing entries that guide the reader

Once the evidence is assembled, the lexicographer decides how to organise the information. Separate numbered entries are usually appropriate when two words have unrelated meanings, different origins or distinct grammatical categories. Shared entries with numbered senses are more suitable when meanings are historically connected and users benefit from seeing the relationship. Cross-references can connect homophones without suggesting that they are semantically related.

Definitions should be concise but discriminating. A weak definition such as “a kind of bank” does little to separate a riverbank from a financial bank. Good examples place the word in a recognisable situation: a customer deposits money at one bank, while a flood reaches the other. Labels such as Australian, informal, technical, archaic or regional help readers interpret the entry without turning the dictionary into a catalogue of unexplained exceptions.

Digital dictionaries add further choices. Search systems may need to return several entries for the same spelling, while voice interfaces must distinguish words that sound alike. Translators and language learners benefit from pronunciation guides, grammatical information and warnings about misleading equivalents. Resources about Japanese manuscript traditions illustrate why form, script and historical context can matter when interpreting written language across cultures; the same principle applies to an electronic dictionary that presents a word outside its original setting.

Applying the method in translation and language teaching

A translator cannot resolve a homonym by selecting the most familiar equivalent. The surrounding sentence, the field of discourse, the intended audience and the relationship between speakers all matter. A legal text may require a precise technical sense, while a novel may deliberately exploit ambiguity. Translation memory systems can preserve an incorrect choice if a human reviewer fails to examine the original context.

This is where false friends and homonyms intersect. Two words may look or sound similar across languages while expressing different concepts, and a translator who relies on form rather than meaning can produce a fluent but inaccurate sentence. The discussion of false friends in translation is relevant because lexical resemblance must always be tested against usage, grammar and cultural context.

Teaching materials should make the decision process visible. Learners can compare example sentences, identify grammatical clues and build small semantic maps rather than memorising disconnected definitions. In Australia, this approach is valuable in classrooms where students may use Australian English alongside a heritage language, an Aboriginal language or another community language. Training and study pathways described through language education resources can support the broader skills required for corpus analysis, dictionary writing and professional translation.

A dictionary also serves communities whose languages have been underrepresented in publishing and technology. Recording homonyms carefully can preserve distinctions that may disappear when a language is forced into an English-based spelling or translation system. In remote communities, consultation should involve fluent speakers, educators and cultural authorities, with consent and control over how recordings and examples are stored. In the Australian language market, this ethical practice is as important as technical accuracy.

A sound lexicographic method therefore moves from observation to classification, evidence, historical comparison and reader-focused presentation. It distinguishes homonymy from polysemy without pretending that every boundary is simple. It uses corpora and fieldwork together, treats local varieties as evidence rather than errors, and gives translators enough grammatical and cultural information to choose the right sense.

The practical takeaway is straightforward: when two words share a form, document their pronunciation, grammar, meanings, examples, history and regional use before deciding whether they belong in one entry or two.