We define that statements are objects that fullfill a certain syntax. But this definition itself is a statement. It is a variant of saying: If the object fulfils a certain syntax then it is a statement because it also fulfils a certain syntax. This is again a statement about the statement and so on...
I know this is very informal and I have started to read a book about logic but the author said he will talk about this subject first at chapter 7 and I hoped that I maybe could get an informal but still comprehensible "peek" of the problem: That we are defining an object in a lower language that still is applicable to higher languages. To me it seems like proving something for fields and then saying that it also holds for rings.
See An Elementary Latin Grammar :
We have here the Latin language, which is the object of the study; call it : object language.
And we have the English language, used to perform the study; call it : meta-language.
The statement "Cesar scribit" is a statement in the object language.
The statement "The statement "Cesar scribit" [means] Caesar is writing." is a statement of the meta-language that expresses a fact about the object language statement "Cesar scribit".
And now compare with : D.van Dalen, Logic and Structure, page 7 :
It is a statement in the meta-language : the usual mathematical argot, made of natural language plus symbols used as abbreviations, regarding the syntax of the object language : the language of propositional calculus.
Yes; it is a statement in the meta-language defining the formal syntax of the objcet language.