The stars around <code>*bug*</code> are called "earmuffs" and are not a special syntax. Rather, they are part of the symbol's name. This widely-used convention effectively places dynamic variables into their own namespace. If dynamic variables are not in their own namespace, thenwhich variablesis intendednecessary tofor bepreventing lexicalbugs. mayCommon accidentallyLisp clashitself withfollows dynamicthis variables of the same name. This is a problemtradition in twoits ways: thestandard dynamic variablevariables is accidentally overridden which can change the behavior oflike <code>*print-circle*</code which depends on its value; furthermore>, this variable is captured by lexical closures as the program expects: it refers to whatever dynamic value is in effect at the time the closure is called, and not the binding which existed when the closure<code>*readtable*</code> waset madecetera.
For similar reasons, earmuffs were used in earlier dialects of Lisp which had only dynamic variables. (One noteworthy surviving descendant dialect is Emacs Lisp.) Variables with earmuffs are intended to be global and occur as free references in code. Variables not denoted by earmuffs are intended to be local. They are always bound in the scope where they are used so their references do not escape (which could happen accidentally, and is usually a bug). An interpreter or compiler for such a dialect could implement checks to enforce this convention.
Common Lisp follows this tradition in its standard dynamic variables like <code>*print-circle*</code>, <code>*readtable*</code> et cetera. Furthermore, the idea of a pervasively special symbol further provides support for the programming style of the traditional dynamically scoped dialects. Programs from such a dialect can be ported to Common Lisp with little changes to how the variables are handled. The "earmuff" variables just have to be properly defined, and most non-earmuff variables can become lexicals.