100% Free Forever
AI-Powered Learning
Industry Expert Content
Certificates & Badges
Learn At Your Own Pace
Programming

Dynamic vs Lexical Scope

How lexical scope resolves variables by source-text structure while dynamic scope resolves them by the call stack, and how Common Lisp uses both via special variables.

Macros & MetaprogrammingAdvanced9 min readJul 10, 2026
Analogies

Two Ways to Resolve a Free Variable

When code refers to a variable that isn't a parameter or local binding, the language must decide how to resolve it: lexical scope looks at where the code was written, walking outward through the enclosing blocks in the source text, while dynamic scope looks at the call stack at runtime, using whichever binding is currently active regardless of where the referencing code physically sits in the source. Modern Common Lisp is lexically scoped by default, but it retains dynamic scope for a special category of variables, making it one of the few mainstream languages where you can observe both scoping disciplines side by side.

🏏

Cricket analogy: Lexical scope is like a fielding position assigned by the written team sheet before the match, fixed regardless of who's captaining live; dynamic scope is like a field placement decided fresh by whichever captain is currently in charge at that exact moment of play.

Lexical Scope in Modern LISP

In lexical scope, a lambda or nested function resolves a free variable by looking at the let/defun/lambda forms textually enclosing it at the point it was written, and this is exactly what makes closures work: the lambda always finds the same binding, no matter who calls it or from where. (let ((y 10)) (lambda (x) (+ x y))) will always add the y bound in that specific let, regardless of what other y might happen to be dynamically active somewhere up the call stack when the lambda is eventually invoked.

🏏

Cricket analogy: This is like a player's registered batting position on the official team sheet always applying to them personally, no matter which other match or team is currently being played elsewhere on the same day.

lisp
;; Lexical scope: the lambda always uses the y from where it was defined
(defun make-adder ()
  (let ((y 10))
    (lambda (x) (+ x y))))

(funcall (make-adder) 5)   ; => 15, always, regardless of caller context

Dynamic Scope with Special Variables

Common Lisp special variables, declared with defvar or defparameter, and by convention named with *earmuffs*, behave dynamically: a let that rebinds a special variable creates a new binding that is visible to every function called during the dynamic extent of that let, not just to code textually inside it. *standard-output* is the classic example: (let ((*standard-output* my-stream)) (some-function-that-prints)) redirects output from any function called inside that let, even one defined in a totally different file, because the lookup happens by call-time context rather than lexical position.

🏏

Cricket analogy: This is like a temporary change of match ball declared for the current innings, every bowler who comes on to bowl during that innings uses the new ball, regardless of which team or coaching manual they personally follow, because the rule applies by current innings context, not by team assignment.

Dynamic scope is genuinely useful for exactly this kind of temporary, call-stack-wide override: redirecting *standard-output* to capture output into a string, temporarily relaxing *read-default-float-format*, or rebinding *print-pretty* for one block of code, all without having to thread a parameter through every intervening function call.

Choosing Between Them

The danger of dynamic scope is 'spooky action at a distance': a function's behavior can change based on bindings set up far away in the call stack, by code the function's author never saw and couldn't have anticipated, which makes dynamic-scoped code harder to reason about locally. This is precisely why Common Lisp restricts dynamic scope to variables explicitly declared special (via defvar/defparameter, or a local declare special) rather than making it the default for every variable, an ordinary let-bound variable is always lexical, so accidentally shadowing it can't silently change some unrelated function's behavior three calls up the stack.

🏏

Cricket analogy: This is like a substitute's on-field decisions being silently affected by a rule change made by team management hours earlier that the substitute was never told about, behavior shifts because of a distant, invisible cause rather than anything visible on the field.

Don't rely on dynamic scope by accident: only variables explicitly declared special (defvar, defparameter, or a local declare special) get dynamic behavior in Common Lisp. An ordinary lexical variable will never be affected by a same-named let binding somewhere else on the call stack — if you need that call-stack-wide override behavior, declare the variable special on purpose, and document why.

  • Lexical scope resolves a free variable by where the code was written; dynamic scope resolves it by the call stack at the moment of execution.
  • Common Lisp is lexically scoped by default — this is what makes closures reliably capture the bindings visible at their definition site.
  • Special variables (defvar/defparameter, *earmuffs* convention) are dynamically scoped: a let-rebinding is visible to every function called during that let's dynamic extent.
  • *standard-output* and similar global-ish variables are the classic legitimate use case for dynamic scope: temporary, call-stack-wide overrides without threading a parameter through every call.
  • Dynamic scope risks 'spooky action at a distance,' where a function's behavior is silently altered by bindings set up elsewhere in the call stack.
  • Because only explicitly declared special variables get dynamic behavior, ordinary lexical variables are safe from this kind of distant interference.

Practice what you learned

Was this page helpful?

Topics covered

#Programming#LISPStudyNotes#DynamicVsLexicalScope#Dynamic#Lexical#Scope#Two#StudyNotes#SkillVeris#ExamPrep