A Practical Guide To Sorting Out Word Categories In English Sentences

Students always get tripped up by words that seem to wear more than one hat. The word "run" shows up as a verb in "I run every morning," but becomes a noun in "I went for a run." That shift isn't random — it's governed by the slot the word occupies in the sentence and what other words sit around it. If you want to stop second-guessing yourself when classifying words, you need to understand how distribution works, not just memorize a list of ten categories. The traditional eight-word-class system covers nouns, verbs, adjectives, adverbs, pronouns, prepositions, conjunctions, and interjections. Interjections like "ouch" or "wow" barely participate in sentence structure, so most people drop them from serious analysis. What remains gives you enough coverage for 95 percent of everyday parsing work. The catch is that these categories describe function, not inherent meaning. A word doesn't carry a fixed part of speech label — it gets assigned one based on its position and behavior within a given context. This is the mistake I see most often. People look at a dictionary entry that says "dark (adjective)" and assume that settles it. Then they hit a sentence like "the dark of night" and get confused because "dark" sits where a noun belongs. Dictionaries list the most common categories first, but they don't tell you the full range of syntactic distribution. Only the actual sentence environment will tell you what that word is doing right now.

Let me walk through a reliable method. Start by drawing out the basic sentence frame and marking where each word falls. If a word appears after a determiner like "the" or "a" and before another noun, it is functioning as an adjective modifying that noun. If it appears after a subject and carries the main action, it is functioning as a verb. These are distributional tests, not meaning tests. You are looking for structural position, not semantic similarity. Consider the word "light." In "turn on the light," it follows the determiner "the" and occupies the object position, which makes it a noun. In "the light is bright," it sits after a copular verb and before an adjective complement, which still treats it as a noun. But in "light the candle," it appears before a direct object, making it a verb. Same spelling. Three different distributions. The pattern tells you the category, not intuition. Another area where people stumble involves function words versus content words. Nouns, verbs, adjectives, and adverbs are content words — they carry the bulk of the meaning. Prepositions, conjunctions, determiners, and pronouns are function words — they build the relationships between content words. A sentence like "The cat sat on the mat" contains two nouns, one verb, two determiners, and one preposition. If you drop any of the function words, the sentence falls apart or becomes grammatically unacceptable. Function words are less flexible than content words because their slots are tightly constrained by the grammar.

Here is a concrete workflow I use when I encounter genuinely ambiguous material. First, check whether the word can take plural marking or a determiner — that points toward noun behavior. Second, check whether it can take tense marking or a negation adverb — that points toward verb behavior. Third, check whether it can modify another adjective or appear in a comparative construction — that points toward adverb or adjective behavior. When a word passes all three batteries, it is genuinely polymorphous, and you need to decide which reading the surrounding context forces. In "the running of the company," "running" passes the verb battery — you could say "the company is running" — but it also passes the noun battery because it sits after a determiner and takes a genitive complement. The phrase structure itself resolves the tension: "running" is a gerund-noun hybrid that behaves like both depending on the construction. I ran into a stubborn case once while grading student work on relative clauses. A sentence read "the book that I bought was expensive." Several students marked "that" as a noun because it referred to "the book." They were wrong. "That" is a relativizer — a subordinator that links the relative clause to the head noun. It does not function as a pronoun inside the relative clause; it has no syntactic role other than introducing the clause. The actual pronoun slot inside the clause is empty and understood as "the book I bought." This distinction matters when you are labeling words for syntactic analysis. Calling "that" a noun in this context muddles the entire structure of the relative clause. Relative clauses are one of the places where beginner grammarians consistently make errors. Words like "which," "who," "that," and "where" can function as relative pronouns or relative adverbs depending on what they replace inside the clause. "The house where I grew up" uses "where" as a relative adverb replacing a locative complement. "The house which I grew up in" uses "which" as a relative pronoun replacing the object of the preposition. Same meaning, different distribution, different label.

Another common trap involves determiners that look like adjectives. Words like "some," "many," "few," and "all" can function either way. In "some people," "some" is a determiner because it sits before the noun and specifies quantity at the noun phrase level. In "some of the people," "some" is a pronoun because it stands in for a noun phrase on its own. You tell the difference by whether the word is followed directly by a noun or by a prepositional phrase. The real utility of understanding Parts Of Speech Of What comes when you need to parse complex sentences quickly. Academic writing, legal documents, and technical manuals are full of nested clauses and nominalizations that shift words between categories. A nominalization like "the implementation of the policy" turns a verb ("implement") into a noun, which means the original verb's argument structure disappears. You can no longer say "implement the policy quickly" using the nominalized form — "quickly" has nowhere to go. Recognizing this shift helps you read more accurately and catch when authors are obscuring agency through heavy noun-stacking. If you want to practice, start by taking short paragraphs from newspapers or reports and labeling every word by its part of speech in context. Do not rely on your gut. Run the distributional tests. You will find that about 10 percent of words resist easy classification, and those are the ones worth spending time on. The remaining 90 percent will fall into place once you internalize the structural patterns.

A word of caution about this approach. Distributional analysis works well for English and similar languages, but it struggles with language varieties that have richer morphological systems or freer word order. In languages where case markings make grammatical roles explicit, you do not need to rely as heavily on positional tests. Even within English, dialectal variation sometimes shifts how words distribute. Informal speech regularly treats "y'all" as both a pronoun and a determiner in ways that formal grammar labels do not capture cleanly. The biggest limitation of the traditional eight-category system is that it was designed for a different kind of linguistic analysis than what modern syntax actually requires. It is useful for basic literacy and introductory grammar courses, but it does not map cleanly onto generative grammar frameworks, dependency parsers, or corpus-based distribution studies. If you are doing anything beyond basic sentence diagramming, you will eventually need a more granular tagset. The Penn Treebank tagset, for example, breaks down "noun" into singular and plural, distinguishes proper nouns from common nouns, and separates adjectives into attributive and predicative uses. That level of detail is overkill for everyday purposes but essential for computational work. For most people reading this, the practical takeaway is straightforward. Learn to identify words by their distribution, not their dictionary entry. Test nouns with determiners and plural markers. Test verbs with tense and negation. Test adjectives with modification and comparison. Test adverbs with sentence position and scope. When a word fails a test, that failure is information — it tells you which category it does not belong to, which narrows the possibilities. Ambiguous words are normal, not a problem to be solved. They simply reflect the flexibility of natural language.

If you want to go deeper, pick a single sentence from a text you are reading and label every word. Then try substituting each word with a known member of its category. If the sentence remains grammatical, your label is likely correct. If the sentence breaks, your label is wrong, and you need to reconsider the distributional evidence. This substitution test is one of the oldest and most reliable methods in the field, and it works across a wide range of sentence types without requiring specialized equipment or software. There is also a practical application in editing and writing. When you understand what part of speech a word is functioning as in context, you can make better choices about word selection, parallelism, and sentence rhythm. Swapping a noun for a verb or an adjective for an adverb is not just a style preference — it changes the grammatical structure of the sentence, which affects clarity, emphasis, and readability. Professional writers use this instinctively. Grammar-aware writers use it deliberately. One final note about the word "what" itself, since it appears frequently in the query this topic stems from. "What" is a wh-word that functions as an interrogative pronoun, an interrogative determiner, or a relative pronoun depending on context. In "What did you buy?" it is a pronoun standing in for the object. In "What book did you buy?" it is a determiner modifying "book." In "What I bought was expensive" it is a pronoun introducing a free relative clause. The same word, three distinct syntactic roles, determined entirely by what sits next to it. This is exactly the kind of pattern that makes distributional analysis useful — it gives you a systematic way to resolve ambiguity without guessing.