Speech errors in Ancient Rome

Question

What records are there of Latin speech errors in Ancient Rome?

I know of spelling errors, e.g. in graffiti, which provide evidence of sloppy or varied pronunciation, but I'm interested to hear about speech errors of the sort that provide evidence of how deeper features of the language, especially the grammar, were represented in people's minds.

The paper "To Err is Human; To Study Error-Making is Cognitive Science" by Hofstadter and Moser provides a variety of illustrations of this kind of thing. For example, in English, a waiter says "Party for two?", accidentally combining the common phrases "Party of two?" and "Table for two?" Another: "The French turn down their noses at rosé wine"—confusing the phrases "turn up one's nose", "look down one's nose", and "turn down". Another: "It was pretty upsettling"—this one blending the words "upsetting" and "unsettling" into a nonsense word.

Presumably these speech errors result from simultaneous processes activating the related phrases. Normally one would "outcompete" the others and inhibit them before the signal reached the mouth, but occasionally they combine. The combined elements, and how they combine, say a lot about how the language is organized in the brain: what "slips" most easily—the competing elements—and what doesn't—the structural properties that are preserved even in the error, like the way "upsettling" has the structure of an ordinary English word even though it's not a real word.*

---

I'm driving at grammar, though, so here's a more relevant illustration from English (from this question on ELL):

The thing that scares me the most are the little oranges in its stomach.

This happens a lot in spoken English—sometimes even in written English. The speaker feels a pressure from "thing" to make the verb singular and a pressure from the upcoming word "oranges" to make the verb plural, and fails to sort it out correctly before the word has come out.

So, did anyone record anything like that from native speakers of Latin?

---

Maybe this will shed some light on a confusion that I was recently relieved of here.

*_{BTW, I accidentally typed "real world" instead of "real word" while writing that sentence.}

Answer 1

I'm not sure hypercorrection is what you're looking for, but if not, have a look at Petronius' Satyricon. The work poked fun at a few of the nouveau riche (Trimalchio is chiefly lampooned), including their bad grammar. I only quote from one book, as the night is late and the author gives examples:

Another point, the admirably clever adaptation of the language to the social position and character of the persons speaking, merits a word or two more. While both the general narrative, and the conversation of the educated dramatis personae, Eumolpus for instance, are marked by a high degree of correctness of diction and elegance of phrase, the talk of such characters as Trimalchio and his freedmen friends, Habinnas and the rest, and other uneducated or half-educated persons, is full not merely of vulgarisms and popular words, but of positive blunders and downright bad grammar. These mistakes of course are intentional, and it is only another proof of the lack of humor and want of common sense that often marked the industrious and meritorious scholars, particularly German scholars, of the old school, that some commentators have actually gone out of their way to correct these errors in the text of Petronius. There are hundreds of them; two or three examples must suffice here. Libra rubricata says Trimalchio (Ch. VII.--xlvi), meaning libros rubricatos, "lawbooks," and vetuo "I forbid," while his guests indulge in such glaring solecisms as malus fatus, exhortavit, naufragarunt. The whole of Chapter VII., where Trimalchio's guests converse freely with one another in the temporary absence of their host, and afterwards Trimalchio harangues the company on various subjects, is full of these diverting "bulls."

Answer 2

I've never come across any kind of systematic study of speech errors, but what might be called the locus classicus for this kind of thing is the well-known Catullus LXXXIV, in which the poet mocks Arrius for misplaced aitches:

Chommoda dicebat, si quando commoda vellet
  dicere, et insidias Arrius hinsidias,
et tum mirifice sperabat se esse locutum,
  cum quantum poterat dixerat hinsidias.
credo, sic mater, sic liber avunculus eius.
  sic maternus auus dixerat atque avia.
hoc misso in Syriam requierant omnibus aures
  audibant eadem haec leniter et leviter,
nec sibi postilla metuebant talia verba,
  cum subito affertur nuntius horribilis,
Ionios fluctus, postquam illuc Arrius isset,
  Iam non Ionios esse sed Hionios.