If "Herbert has deliberately chosen to put his wit in the service of his faith" (Dundas, 225), he has done so in the full awareness of wit's potential ties with the devil's party. Paronomasia is less "admirably suited" to Herbert's staging of wrestling matches between "the claims of the world and the claims of the spirit," as Dundas asserts, than it has been made to suit such a purpose, as she notes a bit further on: "He has deliberately chosen to put his wit in the service of his faith." Paranomasia and its looser forms (ambiguity, pun, verbal slippage, phonetic odd bedfellows, errant ear−events) are not inherently good for enforcing faith; Herbert fights sophistry with sophistry. He chooses his weapons well but not because his weapons of wit were forged expressly to defend faith. In Eleanor Cook's terms, this is Herbert's way of 'troping the scheme' of paronomasia.

Herbert gives this device a further twist. As Dundas points out, "It is a regular feature of Herbert's style to correct one word by another, similar in sound, but more exact from the religious point of view" (227). It is an equally regular feature of his style to represent this correction as a slip of the tongue: "Or if such deceits there be, ⁄ Such delights, I meant to say." To Dundas' observation that here Herbert "proceeds to substitute the word 'deceits' for the word 'delights'" (230), one needs only to add that he "proceeds" to do so through the fiction of a speaker's unpremeditated slip, as though in verbal gaffes we staged a debate between the godly and the worldly. By leaving both the mis−spoken word and the corrected one on the page, the poem suggests an alternative to the silent erasures of written revision. If the poet "often blotted what I had begun" ("Jordan [II]"), the talker has to let both blot and correction stand witness to what he "meant to say." If we only listened to ourselves mis−speak, we would learn the deceit lurking in delight. Perhaps this is one way of resolving the paradox of wit being put at the service of what Dundas calls Herbert's characteristic "devotional simplicity." If a simple, everyday slip of the tongue can be enough of a text on which to meditate the claims of two worlds, who needs a sermon to set text against text? Paronomasia is its own patristics. One can battle with belief in the simplest offhand utterance. The very fact of speaking a non−utopian language, where the same sounds do double duty for often very different meanings, itself becomes an arena for testing faith. As Dundas notes in concluding, Puttenham's "ornament" has in Herbert "the appearance of everyday speech": in part for that reason ornament becomes argument bearing "the force of revelation" (231).

In one paronomasiac scheme, "delights" slips into "deceits" ("The Rose"). In another, "delight" leaves an echoic residue of "light" ("Heaven"). This kind of revelation through reverberation can be difficult to tell from truth through truncation (as in Herbert's "Paradise" with its tercets generated by pruning). How does the retort Word B offers to its paronomasiac pair Word A become modulated through the phonetic changes whereby A may generate, occasion, or lapse into B? The lame tactic of dismissing a claim by a belittling stutter or lisping iteration of one of its words is probably as old as language and is still with us (word, schmerd). To figure such bullying by babbling as pregnant with divine truth is Herbert's achievement.

Dundas also notes that a poem like "The Rose" is to some degree generated from the paronomasiac pairs whose implications it mines, "so important to the very invention of the poem is the pair 'oppose'⁄'rose'" that it "appear[s] almost to set aside the logic of rebuttal" (230). From the most illogical seeds grows the tree of faith—the point is that paronomasia isn't logical, it defies logic, it might well be taken to argue for the vanity of logic itself. How can one reason in a world where "raise" sounds just like "raze"? One cannot reason, perhaps, but one can write poems—these punning pairs almost write the poem for you. Well, actually, they do nothing of the kind; Herbert makes it seem that way, which is what we mean when we say the paired words are the core of the poem's "invention." There's something of the schoolboy exercise about Herbert's poetry (Write a poem that shows how God can "raise" men by "razing" them).

Dundas' provocative article implies that a fuller taxonomy of the ways in which tongue−tinkerings and letter−shufflings enable word to rebut word will help us understand why Herbert had such faith in what paronomasia knew.