Derivational morphology studies the processes by which new words are created from existing morphemes, typically changing the meaning, the lexical category, or both. Affixation is the most common strategy — prefixes (un-, re-, pre-) and suffixes (-ness, -ize, -able) attach to bases to form new lexemes. Compounding combines two or more free morphemes into a single word (blackboard, earthquake), while conversion (or zero-derivation) shifts a word's category without any overt marking (to email from email, to Google from Google). Derivational processes vary in productivity — some affixes actively generate new words (un-Xable is highly productive), while blocking prevents a derivation when an established form already fills the semantic niche (the existence of "thief" blocks "*stealer" in standard usage).
Take a complex derived word like "unbelievability" and peel it apart layer by layer, identifying the base and each affix in the order it was attached. Test the productivity of affixes by trying to coin new words (un-X, re-X, X-ize) and noticing which combinations feel natural and which feel strange. Compare English derivational morphology with a polysynthetic language like Mohawk to see how dramatically derivational capacity can vary.
From your prerequisite study of morpheme types and morphological structure, you know that words are composed of morphemes — the smallest meaningful units — and that these morphemes can be free (capable of standing alone) or bound (requiring attachment to another morpheme). Derivational morphology asks: how does the language engine *create new words* by manipulating these building blocks? The answer involves three core processes, each exploiting different properties of the morphological system.
Affixation is the most visible process. A prefix attaches before the base (un-happy, re-write, pre-heat) and typically shifts meaning without changing grammatical category — "un-happy" is still an adjective, just negated. A suffix attaches after the base and often does change the category: the adjective "happy" becomes the noun "happiness" (adding -ness), or the noun "terror" becomes a verb "terrorize" (adding -ize). This category-changing property is the defining feature of derivational affixes — they derive a new *lexeme*, a word with its own place in the mental lexicon, its own meaning, and its own syntactic distribution. You can stack derivations: "beauty" (noun) → "beautiful" (adjective, -ful) → "beautify" (verb, -ify) → "beautification" (noun, -ation). Each step creates a new word via a new derivational rule.
Compounding builds words from two or more free morphemes: "black" + "board" = "blackboard," "sun" + "flower" = "sunflower," "police" + "man" = "policeman." The compound is a new lexeme with its own meaning — crucially, a meaning that is often not fully predictable from the parts. A "blackboard" need not be black; a "deadline" has nothing to do with death in most contexts. This semantic opacity is the signature of a compound that has been lexicalized — stored as a unit in the mental lexicon rather than computed fresh each time. Conversion (or zero-derivation) takes the process to its limit: the word shifts category without any overt marking at all. "Email" was a noun; now it's also a verb ("I'll email you"). "Google" became a verb ("Just Google it"). The word has gained new syntactic distribution without any morphological change — the derivation is phonologically empty but grammatically real.
Productivity is the measure of how freely an affix generates new words. The affix -able is highly productive: given any transitive verb, you can plausibly coin a new -able word (printable, searchable, crashable). The affix -th (as in warmth, strength, depth) is completely fossilized — you cannot coin new -th words with any expectation they'll be accepted. Blocking is the phenomenon that limits productivity: an established word occupying a semantic slot prevents a derived competitor from filling it. "Theft" blocks "*stealage"; "singer" might block "*songster." The mental lexicon acts like a real-estate market — once a slot is claimed, the derivational process that would create a competitor is suppressed, even though it remains grammatically possible. Blocking and productivity together explain why derivational morphology feels partly systematic and partly arbitrary: the rules are productive, but they operate against a backdrop of established vocabulary that redirects and constrains them.