dissenting.
I agree with my colleagues that the language of 18 U.S.C. § 3146(b)(1)(A), read literally, appears to validate the sentencing rationale employed by the district court. But appearances can be deceiving, and the majority’s result is ultimately both counter-intuitive and at odds with the tiered approach set out in the statute as a whole. See id. § 3146(b).
Judges are not mere grammarians, and the responsibility for statutory interpretation involves more than the application of hard-and-fast syntactical rules to isolated words and phrases. Thus, even plain meaning may succumb to some other, more realistic interpretation if that is necessary to effectuate Congress’s discernible intent. See Holy Trinity Church v. Unit*35ed States, 143 U.S. 457, 459, 12 S.Ct. 511, 36 L.Ed. 226 (1892) (“It is a familiar rule that a thing may be within the letter of the statute and yet not within the statute, because not within its spirit nor within the intention of its makers.”).
In construing statutes, courts have some freedom to depart from literal meaning if doing so is necessary to preserve the statutory purpose. C.I.R. v. Brown, 380 U.S. 563, 571, 85 S.Ct. 1162, 14 L.Ed.2d 75 (1965). As Justice Rehnquist once put it, “in rare cases the literal application of a statute will produce a result demonstrably at odds with the intentions of its drafters.” Griffin v. Oceanic Contractors, Inc., 458 U.S. 564, 571, 102 S.Ct. 3245, 73 L.Ed.2d 973 (1982). When that occurs, “those intentions must be controlling.” Id. The trick, of course, is to recognize those rare cases.8
Given the structure of the statutory scheme, I have come to conclude that this is one of them. Congress’s intent was plainly to create a hierarchy of penalties geared to the magnitude of the sentence that was at stake in connection with the proceeding for which the defendant failed to appear. Had Congress actually considered the application of this statute to failures to appear at supervised release revocation hearings — which I doubt — I believe that it would have written the statute as the defendant would have us construe it.
Although the majority engages in a meticulous dissection of the vocabulary of the statute, the result of that exercise verges on the Kafkaesque. In this instance, I find such slavish devotion to literalism unwise and unwarranted.
The majority’s reading of 18 U.S.C. § 3146 subjects previously convicted individuals to penalties that bear no relationship to the non-appearance of which they are presently accused. As this case illustrates, that approach creates perverse outcomes. The majority’s construction of the statute could expose someone like Smith, who faces a maximum sentence of nineteen months for revocation of his supervised release term, to ten years imprisonment for a simple failure to appear at his revocation hearing. Such a grossly disproportionate outcome is at variance with the sentencing decision reached by the original sentencing judge, who sentenced Smith to time served and supervised release on the underlying conspiracy charge, presumably based on the judge’s own complex determination of aggravating and mitigating factors. We should not presume that Congress intended such an inequitable result. Cf. United States v. Councilman, 418 F.3d 67, 83 (1st Cir.2005) (en banc) (suggesting courts can apply the rule of lenity if, “after seizing everything from which aid can be derived, [a court] can make no more than a guess as to what Congress intended”) (quoting Reno v. Koray, 515 U.S. 50, 65, 115 S.Ct. 2021, 132 L.Ed.2d 46 (1995)).
Ignoring the structure and purpose of the statute, the majority insists that Congress must mean what the statute literally appears to say. To justify this wooden approach, it notes that the statute correlates the severity of the penalty with the seriousness of the charge “in connection with” which the individual “was released.” Ante at 32-33. From this, it leaps to the conclusion that Congress’s intent must *36have been to apply the statute to post-conviction, post-sentencing supervised release revocation hearings and to affix the penalty with reference to the offense that originally led to the supervised release. See id. This conclusion flouts the statute’s evident purpose. It also rests on shaky ground as a matter of statutory construction.
It is a familial' canon of construction that words in a statute are to be interpreted in accordance with the company that they keep. See, e.g., James v. United States, - U.S. -, 127 S.Ct. 1586, 1605, 167 L.Ed.2d 532 (2007). Application of the phrase “released in connection with a charge of’ therefore must be read in context.
Section (b)(1)(A) applies to failures to appear at hearings “if the person was released in connection with a charge of, or while awaiting sentence, surrender for service of sentence, or appeal or certiorari after conviction for [a particular offense].” The listing is chronological and the language, read literally, appears to apply only to failures to appear in proceedings occurring from the time that charges are brought through direct appeal or certiorari after conviction. There is no suggestion that the statute applies to proceedings that post-date conviction, sentencing, and appellate review. Under those circumstances, elevating literalism over practicalities and using the underlying offense to shape punishment in hearings unrelated to the original charge seems inconsistent with the precisely configured time frame of section (b)(1)(A).9
The majority further chides me for focusing on “the proverbial worst case scenario.” Ante at 32-33. It argues that under United States v. Booker, 543 U.S. 220, 261-62, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), district judges retain flexibility to impose below-guidelines sentences, thus foreclosing any concern for grossly disproportionate outcomes. This argument turns a blind eye to the likelihood that the majority’s reading of the statute will make below-the-range sentences (like the one in the instant case) the norm rather than the exception. Something is quite wrong with that picture; the very fact that, under the majority’s construction, a guidelines sentence will routinely result in a grossly disproportionate outcome renders that construction suspect. Cf. United States v. Rita, -U.S.-, 127 S.Ct. 2456, 2464-65, 168 L.Ed.2d 203 (2007) (“[I]t is fair to assume that the Guidelines, insofar as practicable, reflect a rough approximation of sentences that might achieve [18 U.S.C.] § 3553(a)’s objectives.”).
Judge Learned Hand once wrote that, when construing statutes, “it is one of the surest indexes of a mature and developed jurisprudence not to make a fortress out of the dictionary.” Cabell v. Markham, 148 F.2d 737, 739 (2d Cir.1945). I fear that today’s decision runs contrary to that wise advice and erects just such a semantic fortress. Left to my own devices, I would vacate Smith’s sentence and remand for resentencing in accordance with what I perceive to be Congress’s vividly clear intent.
I respectfully dissent.
. The majority, ante at 33 n. 5, cites United States v. Wiltberger, 5 Wheat. 76, 18 U.S. 76, 5 L.Ed. 37, (1820), for the proposition that "where there is no ambiguity in the words [of a criminal statute], there is no room for construction.” Id. at 95. Their reliance on Wilt-berger strikes me as perverse. In that case, Chief Justice Marshall conceded that "though penal laws are to be construed strictly, they are not to be construed so strictly as to defeat the obvious intention of the legislature.” Id. He added that courts can depart from plain meaning in particularly strong cases. Id. at 96. That is precisely my point.
. The majority's charge that this interpretation would materially rewrite the statute, ante at 34 n. 6, is flatly wrong. It would only read the critical phrase — "if a person was released in connection with a charge of” — narrowly, so that it would not apply to proceedings that post-date conviction, sentencing, and direct appellate review. This reading is preferable not only because it leads to more sensible results but also because it is more consistent with the statute’s structure and purpose.