Word Families and Derivation

Learning outcomes

You know the verb “decide.” Now a sentence needs the noun, and you reach for “decidence” or “decidation” and stop, because something tells you neither is real. The word you want is “decision,” and you half-knew it, but the verb did not hand it to you. This happens all day: you own a root in one part of speech and freeze when you need it in another. The fix is not to learn four separate words. It is to learn one family and the rules that connect its members, so that knowing the verb almost gives you the noun, the adjective, and the adverb for free.

After studying this page, you can:

• Define a word family as the set of words built from one root by adding affixes, and turn one known root into its noun, verb, adjective, and adverb forms.
• Match the common suffixes to the part of speech each one creates, so you can build a needed form instead of guessing.
• Choose the right negative prefix for a given root, and recognize that the choice is fixed by the root, not free.
• Explain why the stress often moves when the form changes, and route the detail to your work on stress.
• Use nominalization as a deliberate style choice, knowing when turning a verb into a noun helps and when it buries the meaning.
• Spot the traps: empty slots, irregular forms, and meanings that drift, so a guessed form does not betray you.

Before we dive in

Picture a status update you are writing under time pressure. You want to say that the team can adjust a setting, that this property of the system is worth naming, and that the change went out cleanly. The verbs come easily: “configure,” “scale,” “deploy.” Then the nouns and adjectives stall you. Is the property “scalability” or “scaleability”? Is the setting “configurable” or “configureable”? Did you “make a deployment” or “do a deploy”? Every root you need is one you already command in at least one form. What you are missing is the bridge from that form to the others.

That bridge is derivation: building a new word from an existing one by adding an affix, a small piece attached to the front (a prefix) or the back (a suffix). “Configure” plus the suffix that builds nouns of process gives “configuration.” “Scale” plus the suffix that builds nouns of quality gives “scalability.” The set of all words grown from a single root this way is a word family: “configure, configuration, configurable, reconfigure” are one family, all sharing the root “configure.” The payoff is steep. A learner who treats vocabulary as a list of separate words must meet and store each of those four. A learner who sees the family stores one root and a handful of suffixes, and four usable words fall out. This page teaches that move: how the families are shaped, which suffix builds which part of speech, which prefix flips a meaning, where the pattern is safe to extend, and where it quietly fails.

The family: one root, four jobs

Start with the heart of the idea. English needs the same meaning in different grammatical jobs. You need “decide” to sit as the action in a sentence (a verb), as the thing itself (a noun), as a description (an adjective), and as the manner of an action (an adverb). Rather than invent four unrelated words, English grows them from one root by changing the ending. The root carries the meaning; the suffix assigns the job.

Take “decide.” The verb is “decide” (“we decide today”). Add the noun suffix and you get “decision” (“the decision is final”). Add the adjective suffix and you get “decisive” (“a decisive leader”). Add the adverb suffix to that adjective and you get “decisively” (“she acted decisively”). One root, four forms, four jobs, and the only thing that changed each time was the ending. Notice that the adverb is usually built from the adjective, not from the root directly: “decisive” gives “decisively.” That layering, root to adjective to adverb, is the normal route, and it is worth seeing across several families at once.

Read the table as a grid of jobs, not a list of words. The “decide” row fills all four slots cleanly, which is the ideal case: from the single verb you can reach the noun “decision,” the adjective “decisive,” and the adverb “decisively.” The “analyse” row fills all four too, but watch how far the forms drift: “analyse” becomes “analysis” (the ending changes shape), then “analytical,” then “analytically.” The root is recognizable but the surface has shifted at every step, which is your first warning that derivation is not pure addition. And several rows leave the adverb slot empty: there is no everyday adverb for “configure,” “deploy,” or “scale,” because the language simply never built one. That empty slot is the central trap, and we return to it; for now, hold the win. Learn “deploy” and the family hands you “deployment” and “deployable” with almost no extra effort.

The suffix-to-part-of-speech map

The reason the grid works is that suffixes are not random decorations. Each common suffix is committed to building one part of speech. “-tion” builds a noun, every time. “-ize” builds a verb. “-able” builds an adjective. “-ly” builds an adverb. Once you know which family of endings builds which job, you stop guessing and start constructing: you need a noun from “implement,” you reach for a noun-building suffix, and “implementation” appears. The map below is the engine of the whole skill.

Read the map as a set of tools, each shaped for one job. The noun row is the busiest because English builds nouns in many ways: “-tion” and “-sion” mark a process or its result (“implementation,” “decision”), “-ment” marks an action or its outcome (“deployment”), “-ness” turns an adjective into the quality of being it (“awareness”), “-ity” does the same for more formal adjectives (“scalability”), “-ance” and “-ence” mark a state or action (“performance,” “preference”), and “-er” or “-or” name the person or thing that does it (“deployer,” “operator”). The verb row is smaller and more pointed: “-ize” and “-ify” mean roughly “make into” or “make more” (“modernize,” “simplify”), “-en” does the same for short adjectives (“shorten,” “harden”), and “-ate” marks an action (“validate,” “automate”). The adjective row gives you a description from a noun or verb: “-able” and “-ible” mean “can be done to” (“configurable,” “reversible”), the most productive of all for a software learner, since almost any action verb yields an “-able” adjective. And “-ly” is the one reliable adverb maker: attach it to almost any adjective and you have the adverb. Knowing this map, you no longer ask “what is the noun for implement?” You ask “which noun suffix does this root take?” and the answer is usually one of six.

Negative prefixes and which root takes which

So far every affix has been a suffix on the back of the root, changing the job. Prefixes work on the front and usually change the meaning instead. The most useful group is the negative prefixes, the ones that flip a word to its opposite: “un-,” “in-” (with its spelling variants “im-,” “ir-,” “il-”), “dis-,” “non-,” and “mis-.” The trap is that you cannot pick freely. Each root has already been assigned its negative prefix by the history of the language, and the wrong one is simply not a word. “Unpossible” and “impossible” are not two styles of the same thing; only “impossible” exists.

A few of the choices follow a sound rule. The “in-” prefix changes its spelling to match the first letter of the root: it becomes “im-” before “m” or “p” (“impossible,” “immature”), “ir-” before “r” (“irregular,” “irreversible”), and “il-” before “l” (“illegal,” “illogical”). That assimilation is predictable, so those three you can build by ear. But the larger choice, whether a root takes “un-,” “in-,” “dis-,” “non-,” or “mis-” in the first place, is mostly fixed and must be learned per word. The flowchart below captures the part you can rule, and names the part you cannot.

Read the flowchart top to bottom as two different kinds of decision. The lower-right branch, once you know a root takes “in-,” is fully ruleable: the prefix just reshapes itself to the first letter, giving “impossible,” “irregular,” “illegal,” and plain “invalid” everywhere else. That branch you never have to memorize word by word. But the very first question, the top branch, has no rule behind it. Why “unhappy” and not “inhappy,” why “impossible” and not “unpossible,” why “disagree” and not “unagree,” why “non-negotiable” and not “innegotiable”: these are settled by the root’s history, often by whether the root came into English from Latin (which favors “in-”) or from older Germanic stock (which favors “un-”). The practical takeaway is to learn the negative form as part of the family, the same way you learn the noun form, rather than trying to derive it. The senses differ slightly too: “dis-” often undoes an action (“disagree,” “disconnect”), “mis-” means done wrongly (“misunderstand,” “misconfigure”), and “non-” is a flat, neutral negation (“non-negotiable,” “non-blocking”), which is why a release note says “non-blocking” rather than “unblocking,” a word that means the opposite.

Test the choice on a fresh root before moving on.

The stress moves when the form changes

There is a catch that the spelling hides. When you derive a new form, the stress, the syllable you say with the most force, often jumps to a new place. The family looks like one word with different endings, but it can sound like four different words. This matters because a derived form said with the wrong stress is hard for a listener to recognize, even when the spelling is perfect.

The classic case is the “photograph” family. The noun “photograph” stresses the first syllable: PHOtograph. Add “-y” for the activity and the stress jumps to the second: phoTOgraphy. Add “-ic” for the adjective and it jumps again, to the third: photoGRAPHic. Same root, three forms, three different stressed syllables. The “economy” family does the same: the noun is eCOnomy, but the adjective is ecoNOMic, with the stress pulled one syllable to the right by the “-ic” ending. Certain suffixes, “-ity,” “-ic,” “-ion,” and a few others, reliably tug the stress toward themselves, which is why “scale” (one syllable, no contest) becomes “scalaBILity” with the weight near the end.

The point for now is awareness: when you build a derived form, do not assume it keeps the root’s stress. Say it, and check where the force lands, because moving the stress can change which word a listener hears. The full system of which suffixes move stress and how to place it lives in word stress and clarity, and it is worth studying there, because a correctly built but wrongly stressed word still fails to communicate.

Nominalization: a style choice, not a free upgrade

Once you can turn verbs into nouns at will, a new power appears, and with it a new way to write badly. Turning a verb into a noun is called nominalization: “decide” becomes “the decision,” “implement” becomes “the implementation,” “fail” becomes “the failure.” The grammar lets you take the action of a sentence and package it as a thing. Used well, this adds formality and lets you treat a process as an object you can discuss. Used heavily, it drains the life out of prose, because it hides the action inside a noun and often drops the person doing it.

Compare two versions of the same report sentence. Verbal style: “We decided to roll back the release because the new index slowed every query.” Every action is a verb, every actor is named, and the sentence moves. Nominal style: “The decision to perform a rollback of the release was made due to the degradation of query performance by the new index.” The same content, but now “decided” has become “the decision,” “roll back” has become “a rollback,” “slowed” has become “the degradation,” and the person who decided has vanished into “was made.” The second version is not wrong, and in a formal incident report parts of it might even be wanted, but it reads as heavy and bureaucratic, because the actions are frozen into nouns and the actor is gone.

So nominalization is a dial, not a switch. Turn it up for a formal abstract or a clause in a contract, where naming the process as a thing is exactly the point. Turn it down for anything that should read clearly and move, which is most writing most of the time. The full treatment of how a verbal style sharpens prose, and how to find and unwind a buried action, is the subject of principles of clear prose. For our purposes the lesson is that derivation gives you the power to nominalize, and a skilled writer uses it on purpose, not by default.

Where the pattern breaks

By now the family model is doing a lot of work for you, which is exactly when it is most dangerous, because it tempts you to extend it past where the language actually goes. Three boundaries matter, and each one quietly produces errors that sound confident.

The first boundary is the empty slot. Not every family fills all four jobs, as the very first table showed. “Configure” has no everyday adverb; “deploy” has no common adjective beyond “deployable”; many abstract nouns have no verb at all. The rule “every root has four forms” is false. When a slot is empty, the form you would build for it is a non-word, and reaching for it (“we deployed it deployingly”) marks you instantly. Before you build a form, ask whether the language actually has it, rather than assuming the grid is full.

The second boundary is the irregular form. Derivation is rarely clean addition. “Decide” does not become “decidation”; it becomes “decision,” with the root reshaped. “Analyse” becomes “analysis,” not “analyseation.” “Maintain” becomes “maintenance,” not “maintainment.” The suffix map tells you which job a suffix does, but not the exact spelling of the join, which can warp the root. So the map points you to the right region (“a noun suffix goes here”), and then you confirm the actual form, because the obvious build is often not the real one.

The third boundary is the meaning that drifts. Two forms from one root can split into different meanings, so the form you build is real but says the wrong thing. The sharpest example is “economic” versus “economical.” Both come from “economy,” both are adjectives, and they are not interchangeable: “economic” means relating to the economy (“economic policy”), while “economical” means thrifty, saving money or resources (“an economical car”). Pick the wrong one and you have written a real word that means something you did not intend. Similarly “historic” (momentous) is not “historical” (relating to history), and “a sensible plan” (wise) is not “a sensitive plan” (touchy, easily hurt). When a root has two adjectives, suspect that they have divided the meaning between them.

Read the table as three different failures of the same overconfidence. The empty-slot row fails because the form was never built; the irregular row fails because the form exists but in a reshaped spelling; the meaning-drift row fails because the form is perfectly real but carries a sense you did not want. All three come from trusting the pattern past its edge. The defense is the same in each case: treat a freshly built form as a hypothesis, not a fact, and confirm it, exactly the care you bring to precision vocabulary, where choosing the form that means precisely what you intend is the whole game.

Mental Model: learn the family, not the word

Here is the reframe that turns derivation from a memory burden into a multiplier, and it corrects the model most learners start with. The wrong model is that vocabulary is a flat list of separate words, and that “decide,” “decision,” “decisive,” and “decisively” are four independent entries you have to meet and store four times. Under this model your vocabulary grows one word at a time, and you are forever short of the exact form a sentence needs, because you happened to learn the verb but not the noun.

The better model is that vocabulary is a set of families, each a single root that flexes into several jobs. You do not learn “decide” as a point; you learn the “decide” family as a small cluster, with the verb at the center and the noun, adjective, and adverb arranged around it by their suffixes. When you meet a new root in reading, the productive move is not to confirm you know its one form but to grow the rest: see “deploy,” and immediately note “deployment” and “deployable,” so the family is stored whole. The suffix map is what lets you do this almost for free, and the boundaries, empty slots, irregular joins, drifting meanings, are what stop you from over-growing the family into non-words.

Read the diagram as the unit you should actually store. The center is the root and its meaning, but the value is in the spokes: each form is the root plus the suffix that assigns its job, with the adverb hanging off the adjective and the negative coming from a prefix on the front. Learn this shape once and you carry four or five usable words, and you can see at a glance which jobs the family fills and which it leaves empty. This is the same redirection of effort that register-specific vocabulary asks for at the level of formality: do not collect isolated words, collect the structured family, so the form you need is already attached when the sentence calls for it.

Common mistakes

The first mistake is inventing a non-word by forcing a slot the language never filled. Wanting an adverb for “configure,” a learner writes “configuredly”; wanting a noun for “decide,” they write “decidation.” The form is built from real pieces but does not exist. The fix is to remember that the grid is not always full and the join is not always regular, and to confirm a freshly built form before trusting it, rather than assuming every root yields every form.

The second mistake is the wrong negative prefix: “unpossible,” “inhappy,” “displike,” “unregular.” Each pairs a real root with a real prefix, but the root takes a different prefix, so the result is a non-word. The danger is that negation feels logical, so the wrong prefix feels right. The defense is to learn the negative form as part of the family, and to recall the “in-” spelling rule (“im-” before m or p, “ir-” before r, “il-” before l) only for roots you already know take “in-.”

The third mistake is the wrong suffix for the part of speech you need: building a noun with an adjective ending, or stacking a suffix the root does not take (“performant” where you needed “performance,” “scaleable” where the “e” should drop). The fix is the suffix map: decide which job you need first, reach for a suffix that builds that job, then check the spelling of the join.

The fourth mistake is over-nominalizing: turning every verb into a noun until the prose is a parade of “-tion” and “-ment” words with no one doing anything. “The implementation of the optimization resulted in the elimination of the degradation” is grammatical and almost unreadable. The fix is to treat nominalization as a deliberate, occasional move for formality, and otherwise to keep the action in the verb and the actor in the subject.

The fifth mistake is trusting a built form whose meaning has drifted: writing “economic” for thrifty, “historic” for about history, “sensible” for touchy. The form is real, which is why no spell-checker flags it, but it means something you did not intend. When a root has two derived adjectives, assume they have divided the meaning and check which is which before you commit.

Mastery Questions