This question already has an answer here:

Mathematica Version 11.2 desktop / Windows 10 Pro 64-bit

(* Assume y is undefined. *)

(* ?y and ??y leave y undefined. *)


Information::notfound: Symbol y not found.


Information::notfound: Symbol y not found.

(* AtomQ[y] has a side effect of defining y. *)

In[3]:= AtomQ[y]

Out[3]= True

In[5]:= ??y


Why does AtomQ[y] define y?

Is this "By Design" i.e. a carefully thought out deliberate feature?

One would think that if y is not defined, AtomQ[y] would output this as an error.

Why else would one be using AtomQ[y] unless they thought y was already defined.

It seems like a dangerous side effect which makes debugging a nightmare.

Is there a property of the Kernel to enable or disable functions like AtomQ from defining symbols?

In general, which built-in functions leave symbols undefined and which do not?

Thanks in advance.


marked as duplicate by m_goldberg, bbgodfrey, Kuba Nov 27 '17 at 8:38

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • $\begingroup$ The creation and removal of symbols in contexts is described in the tutorial reference.wolfram.com/language/tutorial/BasicObjects.html. I'm not sure "define" is the right term to describe what happens. For instance Definition[y] returns Null, if you haven't given y any definitions or attributes, even though the name "y" is added to the "Global`" context. $\endgroup$ – Michael E2 Nov 26 '17 at 21:12
  • $\begingroup$ Invoking foo[bar] puts both of foo and bar into the name space, in whatever is the current Context. But it does not mean either one has become "defined' in the sense of having DownValues or the like associated to them. $\endgroup$ – Daniel Lichtblau Nov 26 '17 at 21:31
  • 1
    $\begingroup$ It doesn't define it. It creates the symbol. Simply saying y creates the symbol y. $\endgroup$ – Szabolcs Nov 26 '17 at 22:11
  • $\begingroup$ I think linked topic answers your doubts and provides much more background. Let me know if you disagree with closing. $\endgroup$ – Kuba Nov 27 '17 at 8:39

This has nothing to do with AtomQ. Constructing an expression containing a symbol automatically gives it a definition. It doesn't even need to be evaluated. It acquires a presence in some context (usually Global), but not a value.

"Definition" isn't a formal Mathematica concept. Is a symbol that exists in some context "defined"? Yes, in the sense implied by this question, but no in the sense of having a value.

  • $\begingroup$ @Kuba As the comments above imply, "definition" isn't a formal Mathematica concept. Is a symbol that exists in some context "defined"? Yes, in the OP's sense, but no in the sense of having a value. $\endgroup$ – John Doty Nov 27 '17 at 12:53
  • $\begingroup$ Thanks for the edit $\endgroup$ – Kuba Nov 27 '17 at 15:39

Not the answer you're looking for? Browse other questions tagged or ask your own question.