What was Aristotle's concept of logical form?

This paper argues (contra Timothey Smiley and John Corcoran) that Aristotle did not invent a 'formal language' for his syllogistic, and (contra Jonathan Barnes) that Aristotle did not analyse syllogistic validity in terms of logical form, if a logical form is some sort of syntactical schema. The paper also argues that Aristotle does not hold there is a canonical way of propounding a syllogism: namely, by using locutions such as 'X does/does not belong to all/some Y'. In this he differs from the Stoics, who probably did hold that there was a canonical way of presenting syllogisms. The paper concludes that these unusual locutions which Aristotle uses are ways of generalising over what propositions say, rather than attempts to (i) invent a formal language, (ii) provide a syntactical logical form, or (iii) suggest canonical language in which to propound syllogisms.