Horace, Satires 1.3.117–23, as transmitted:

Horace is attacking the Stoics for their doctrine that all sins are equal. The overall run of sense must be that given above. If kept, the ut of line 120 must then have the value of ne: were it allowed its normal meaning in expressions of fear (‘lest not’) the opposite of what is required would be said. The oddity of the construction was long excused as a sort of anacoluthon, provoked by the postponement of uereor. Such defences of the received text were rightly criticised by Palmer and Housman: ‘incredible’, they both concluded, and indeed it is incredible that Horace could have in such a short space forgotten what he had written, and forgotten it in such a way as to say the opposite of what he meant.Footnote ²

Palmer's own remedy was to replace non by nunc in 121. It is to be rejected, since it makes of excessive leniency a positively desirable thing, which Horace wishes the Stoic would practise but fears he will not. Housman emended more radically:

He understood the ut as concessive, paraphrasing ‘for (though I daresay you will inflict the flagellum on one who deserves only the scutica) I am sure from your talk that you will not be content with inflicting the ferula on one who deserves something worse’. Shackleton Bailey took a different approach:Footnote ³

That is, ‘I don't care if you strike with a schoolmaster's rod one who has earned more serious lashing …’. Of the above rewritings, it is fair to say, none reads smoothly or easily; in addition, Bailey's conjecture supposes a more complex error than one might comfortably wish, Housman's a more unusual one. Non uereor has also been emended to non uincor, giving ‘I am not persuaded that you strike with a schoolmaster's rod one who has earned more serious lashing …’.Footnote ⁴ But the point to be made is made better with uereor: excessive leniency by Stoics is not to be feared, that is to say it is not the danger inherent to their dogma. A solution through punctuation may be mentioned, too: nam ut ferula caedas meritum maiora subire | uerbera? non uereor, cum dicas …Footnote ⁵ But the mocking question in ut should react to a possibility already raised, whereas that of laxity from a Stoic has not been. If forced to choose among proposals to date, we would do least disservice by just replacing nam ut with ne.Footnote ⁶ But we would still be off the mark.

In fact Horace wrote ut, non and uereor; he meant by them non uereor ne; moreover he did so quite deliberately. He was imitating ὡς for μή in clauses of fearing. As in Horace's Latin calque, the Greek idiom appears when the governing expression is negatived.Footnote ⁷ Though never very frequent, οὐ δέδοικα (ϕοβοῦμαι, etc.) ὡς is none the less a regular alternative in Attic to οὐ δέδοικα (ὅπως) μή:

To be sure, writing ut while meaning ne was a cocky move, even for Horace. It was bound to startle. It could also very easily confuse. Accordingly the surrounding passage has been carefully structured: note the nam, which leaves no doubt as to where the thought is going. Postponement of the governing verb spares us uereor followed directly by ut, which would have left the anomaly of construction starker and heightened an impression of contresens. Another strategy, too, is at work to assist the reader. In the masculine singular, a perfect participle such as meritum may not normally stand for hominem meritum, and to use one so is suggestive of Greek.Footnote ⁹ Likewise magnis parua falce simili recidere is not native Latin, which would say instead magna et parua falce simili recidere; the model is the construction common with ὅμοιος, and also for that matter with ἴσος and ὁ αὐτός.Footnote ¹⁰ Note also the cum of line 121. It has most point if taken in a purely temporal, rather than a causative, sense: ‘I'm not afraid that you may be too lenient when you say mere thefts are of a pair with brigandage …’. One therefore expects dicis … et … minaris, not dicas … et … mineris. So why the subjunctives? Assimilation of the cum-clause into the ut-clause may bear some responsibility, but so too may a colouring of cum by ὅταν. In other words, one sentence holds at least two, and quite possibly three, linguistic Hellenisms in addition to ut = ὡς, making for an extraordinary total.Footnote ¹¹ On finding himself in this thicket of alien usage the reader is served in effect the key to the puzzle: to make sense of ut … caedas he must think in Greek.

There may well be a second instance of the construction invented by Horace. Livy 28.22.11–12, describing a surprise sortie by the besieged at the Spanish town of Astapa in 206:

When not emended into regularity,Footnote ¹²nihil minus quam, ut egredi moenibus auderent, timeri poterat has been analysed as something other than an expression of fearing, usually in terms of those noun clauses beginning in ut, difficult to fit with exact grammatical labels, that one tends to explain as meaning ‘the proposition that’, ‘the possiblity that’ etc.Footnote ¹³ So Palmer (n. 2) made the ut ‘epexegetic of illud understood’; Weissenborn saw a construction of expectation.Footnote ¹⁴ A more promising view along such lines refers to the ut-clauses occurring as the second term of a comparison, after quam (so for example nihil uidi tam integrum quam ut oppugnare imperatorem incipiant [‘I have noticed nothing so open to them as to start an attack upon their commander’] at Cic. Phil. 13.34); Livy's expression would then represent this idiom, or a hybrid of it and that normal with verbs of fearing.Footnote ¹⁵ Notwithstanding, it remains possible that Livy had an eye on Horace's daring experiment in Satires 1.3. Note that ut … auderent depends on a negative expression, as in Horace, and that the governing verb, as in Horace, is postponed. Livy allowed himself to be inspired by contemporary lyric and even elegy;Footnote ¹⁶ satire is not to be excluded.