dissenting.
¶59 Amber Lee Torrez spent more than four years confined before serving a sentence for any crime. In the end, the trial court allowed her to credit 86 of those 1579 days toward the sentence it imposed. The majority would afford her even less: If it could consider the- 86 days she did receive, its test would strip her of that credit as well. And more broadly, its holding seems to deny presen-tence confinement credit to every offender confined on multiple charges,
¶60 How does this counterintuitive result emerge? The answer in part is that in purporting to recite our test for ■ determining whether an offender should receive presen-tence confinement credit (PSCC), the majority instead rewrites it. This court previously asked whether an offender confined on multi-*202pie charges would have “remained confined in the same judicial district on the charge for which credit is sought in the absence of any other charge." Massey v. People, 736 P.2d 19, 23 (Colo. 1987) (emphasis added). Now, however, the majority states that an offender cannot receive credit if she “would have remained confined regardless of the sentencing charge.” Maj. op. ¶ 26.
¶61 Some 'simple hypotheticals demonstrate how omitting the place-based language makes a difference. Suppose Jane Doe injures a pedestrian in a hit-and-run, for which Durango issues an arrest warrant against her. A month later, she runs headlong into Durango Police while robbing a convenience store. That encounter places her in the La Plata County jail on both charges, where she stays until she’s convicted and. sentenced. Under our now-discarded test, she would have received credit for that entire period of confinement, Under the new test, because she would have remained confined on one of the charges in the absence of the other, arguably she receives none. •
¶62 Or consider an even likelier scenario'. The People charge Doe with multiple offenses arising' out of a single incident (as was the case for Torrez in Jefferson County). She remains in the same jail throughout proceedings. Doe-eventually pleads guilty to one of the charges, and the court dismisses the rest. Here again, though our former test would have entitled Doe to credit, the majority’s new test seemingly offers her none-r-she would-have remained confined in the absence of the sentencing charge, given the remaining charges^ And if she had instead chosen a trial and lost? The same.
■ ¶63 As I .read the majority’s new'test, then, only offenders whose confinement is solely attributable to the sentencing charge will receive credit. We have never before required such a result, and the presentence confinement statute’s plain language does not compel it. To the contrary, that statute entitles Torrez to credit for her full period of confinement. Therefore, I respectfully dissent.
¶64 In crafting my dissent, I write more expansively than I otherwise might, so that the legislature can perhaps better evaluate whether it should intercede to make its intentions more plain, particularly given the significant implications today’s majority opinion holds for defendants’ liberty interest and the state budget.
I. Analysis
. ¶65 Neither our precedent nor the language of- the current PSCC statute supports the test the majority announces today. Instead, this court has used geography as a proxy to determine whether an offense caused a defendant’s confinement for purposes of section 18-1.3-405, C.R.S. (2017). Even that place-based test, looks suspect, however, given that the legislature has overhauled section 18-1.3-405 since we last explored its requirements. With those changes, the legislature granted credit broadly and narrowed that broad grant where.it might ■prove problematic. Because Torrez’s circumstances do not fall within any of those limitations, and her Jefferson County offense concurrently .caused her confinement, I conclude section 18-1.3-405 entitles her to PSCC for the entire time she spent confined—namely, from the time of her arrest until she was sentenced for the Jefferson County offense.
A. We Have Never Employed the Test the Majority Articulates Today
¶66 The presentence confinement statute has evolved in response to interpretations from this court. I relate that history more fully below and provide examples along the way to illustrate the effects of various changes and interpretations, but my analysis in-this subsection reduces to three points. First, early iterations of the f^SCC statute left significant ambiguities regarding its application. Second, in the absence of legislative guidance, this court created its own framework for awarding credit. Third, from this overview, I conclude not only that we have never employed the test the majority articulates today, but that the new test requires results at odds with those we have approved.
¶67 The original, 1972 PSCC statute merely required a sentencing judge to take pre-sentence confinement “into consideration” but did not mandate an award. Ch. 44, sec. 1, *203§ 39-11-306, 1972 Colo. Sess. Laws 190, 249. It survived only seven years. After .courts consistently construed the 1972 statute to allow discretion in awarding PSCC, the legislature repealed and replaced it in 1979.1 Schubert v. People, 698 P.2d 788, 792-93 (Colo. 1985) (discussing statutory history related to People v. Jones, 176 Colo. 61, 489 P.2d 596, 599-601 (Colo. 1971)).
¶68 The 1979 language mandated PSCC but created new difficulties because it appeared to require credit for any period of presentence confinement, no matter its cause. As-written, the statute eliminated a disincentive to reoffend while confined because it seemed to credit time already served against both new and old sentences. See id. at 796. If, for example, an offender assaulted a prison guard and then spent 100 days confined before being sentenced for the assault, the statute appeared to credit 100 days toward her original sentence as well as 100 days toward her new sentence on the assault charge. Thus, .the assault would go largely unpunished.
¶69 Addressing that anomaly, our court created the “substantial nexus” text. Under that test, we required courts to consider two factors. First, we held that “the presentenc'e confinement [must be] actually caused by the charge or conduct for which the offender is to be sentenced[.]” Id. Second, contrary to the majority’s understanding, we observed that multiple offenses could concurrently cause a defendant’s confinement and added, “[T]here must bé a substantial 'nexus between such charge or conduct and the period of confinement for which credit is sought.” Id. Thus, under the framework we 'announced, the prison assailant from the earlier example would receive 100 days of credit toward her original sentence, but not toward the new one. See id.
¶70 In illustrating how to apply this new test, we went on to suggest an additional consideration: whether, in the event multiple offenses caused the confinement, they originated in the confining jurisdiction. To that end, we-said,
If, for example, multiple counts or cases are concurrently filed against a defendant ia the same jurisdiction, and the- defendant remains confined in that jurisdiction on all charges due to his inability to post bail, each charge would appropriately be considered a cause of the defendant’s presen-tence confinement.... If only one sentence is ultimately imposed and the other concurrently filed counts or cases are dismissed-,, then clearly the defendant should be credited witb the entire period of pre-sentence confinement served in the sentencing jurisdiction against the sentence.
Id. at 795-96 [emphases added) (citation omitted). We further noted that even if the charges were not filed concurrently, the statute still required courts to credit the period of confinement following' the later-filed charge toward the sentence resulting from that charge. Id. at 796 n.13. And, taking the same position in Schubert’s companion case, Torand v. People, 698 P.2d 797, 801 (Colo. 1985), we held that “the pendency of another case in the same jurisdiction [did not] serve to negate the substantial nexus between the charges in the instant cáse and the defendant’s confinement.” (Emphasis added.)
¶71 Under the test we articulated in Schubert and Torand, if an offender were, confined for 100 days in a Denver, jail on a Denver assault charge as well as a Denver burglary charge, she could credit all 100 days toward her sentences for both, or could credit the 100 days against one of the sentences if the other charge were dismissed.
¶72 The majority, on the other hand, seemingly would offer her no credit at -all because she would not “have been releaséd from the confinement had [either the assault charge or the burglary charge] not existed.” See maj. op. ¶ 3. In fact, the majority’s new test *204renders Torand wrongly decided: Under the majority’s test “the pendency of another case in the same jurisdiction” does negate the substantial nexus and destroys an offender’s right to credit.
¶73 Still, the majority also roots its rule in Massey, a case we decided two years after announcing Schubert and Torand. Again construing the 1979 statute, we returned to the substantial nexus test and, far from adopting the majority’s but-for test, we made Schubert’s suggested geographic limitation part of our rule.
¶74 In Massey, a defendant arrested on warrants from Mesa and Pitkin Counties sought credit in Pitkin County for his time spent confined in Mesa County while waiting to resolve the charges from both counties. 736 P.2d at 20. Worried about awarding du-plicative credit, yet finding no statutory guidance on the issue, we looked to Schubert’s illustration and adapted the substantial nexus test to the circumstances. We therefore held, “A ‘substantial nexus’ [where two or more charges form multiple bases for the defendant’s presentence confinement] means that the defendant would have remained confined in the same judicial district on the charge for which credit is sought in the absence of any other charge.” Id. at 23 (emphasis added). Applying that test, we determined the defendant was entitled to PSCC only for the time he spent confined in the jurisdiction issuing the sentence. Id.
¶75 Consider, under this scheme, the example of the multiple offender. This time the offender is charged in two cases from Denver, one for burglary and the other for assault. She spends 100 days confined before beginning to serve her sentences, but of those 100 days, she spends 16 days resolving the assault charge, 30 days resolving her burglary charge, and 66 days waiting for the ■cases to move forward. Under the Massey test, the court must credit her 100 days of confinement toward the sentences for both offenses because she would have remained confined, in the jurisdiction on either charge in the absence of the other.
¶76 Now consider the effect of the same charges if Adams County files the burglary charge. The offender spends 15 days in Denver resolving the Denver assault charge, 30 days in Adams County resolving the burglary charge, and 55 days in Denver waiting for the cases to move forward. Under Massey, the court must credit 70 days toward her sentence on the assault charge, and 30-days toward her sentence on the burglary charge. Furthermore, if the assault charge is dismissed in the Denver-only example, the offender receives credit for all 100 days. If it is dismissed in the Denver/Adams County example, the offender receives only 30 days of credit. When maintaining a causal common denominator, then, geography makes all the difference.
¶77 The majority collapses the geographic distinction, but only by denying the multiple offender credit in both scenarios. In .either scenario, the offender would not have been released had the sentencing offense not existed.
¶78 Despite reciting the words “substantial nexus,” then, it seems the majority has eliminated any opportunity for credit for an offender confined on 'multiple charges. Yet, as Schubert makes clear, we created the substantial nexus test precisely because we recognized that multiple offenses could concurrently cause an offender’s confinement. Not anymore, it seems. And though the majority is of course free to disavow Schubert’s test and to declare that Torand was wrongly decided, stare .decisis at the veiy least requires it to explain why its new test is better suited to the task. See People v. Kutlak, 2016 CO 1, ¶ 18, 364 P.3d 199, 205 (observing that this court departs from an established rule only upon good reason). It has not done so.
¶79 Perhaps, then, the new test emerges from the legislative retrofit that followed those cases? Let’s see.
B. Section 18-1.3-405 Does Not Mandate the Majority’s Test
¶80 The General Assembly has not mandated, as the majority now does, that an offender’s confinement must be solely attributable to the charge for which she seeks credit. On the contrary, nothing in the statute’s plain language prevents a court from *205awarding credit when more than one offense causes a defendant’s confinement.
¶81 Responding to Schubert in 1986, the legislature added its own, broad causal requirement: “A person who is confined for an offense prior to the imposition of sentence for said offense is entitled to credit against the term of his sentence for the entire period of such confinement.” Ch. 124, sec. 3, § 16-11-306, 1986 Colo. Sess. Laws 733, 734 (additions emphasized). Then, in 1988⅛ the General Assembly inserted language addressing the prison-assault scenario that had worried the court in Schubert. Ch. 110, sec. 2, § 16-11-306, 1988 Colo. Sess. Laws 663, 663-64. And in 2009, the legislature inserted language directing courts to credit an offender’s confinement pending a parole revocation hearing toward any period of reincarceration imposed in the parole revocation proceeding. Ch. 105, see. 3, § 18-1.3-405, 2009 Colo. Sess. Laws 382, 383. Thus, the statute now states as follows:
• A person who is confined for an offense prior to the imposition of sentence for said offense is entitled to credit against the term of his or her sentence for the entire period of such confinement. ...
• A person who is confined pending a parole revocation hearing is entitled to credit for the entire period of such confinement against any period of rein-carcération imposed in the parole revocation proceeding. ...
• If a defendant is serving a sentence or is on parole for a previous offense when he or, she commits a new offense and he or she continues to serve the sentence for the previous offense while charges on the new offense are pending, the credit given for presentence confinement under this section shall be granted against, the. sentence the defendant is currently serving for the previous offense and shall not be granted against the sentence for the new offense.
§ 18-1.3-405.
¶82 To determine whether the statute compels or at least supports the majority’s test, I use our standard rules of statutory construction. When a statute’s language is clear, this court must apply it as written. People v. Smith, 2014 CO 10, ¶ 13, 318 P.3d 472, 476 (interpreting section 18-1.3-405).
¶83 We cannot add words to the statute, id., yet the majority does just that. Read naturally, the word “for” doesn’t require exclusivity. If a doctor’s office places a sign in the waiting room that reads, “Those visiting for a condition must cover the cost of consultation for said condition before leaving,”' I very much doubt I could avoid paying just because the doctor attended to both my sore throat and broken toe. Yet this is exactly how the majority interprets the' PSCC statute—not as written, but rather to say, “A person who is confined for an offense and only that offense ... is entitled to credit. ...”
C. Section 18-1.3-405 Grants Credit. Broadly and Confines it Where it May Prove Problematic
¶84 Having concluded the majority’s test finds no support in our precedent or the statute’s text, I consider what the current statute does require. Although I see the allure in simply concluding Massey’s bright-line, place-based test has endured, I.cannot apply the old test to the eurrerit statute without first assessing whether the statute still requires that test. See Specialty Rests. Corp. v. Nelson, 231 P.3d 393, 403 (Colo. 2010). Looking .to the current statute’s plain text, I do not read it to require such dispar rate treatment based on geography alone.
¶85 This court-created the geographic limitation in the absence of any specific instructions from the legislature. The legislature has since supplied those "instructions. With its 1986 amendment, the legislature identified the causal requirement we thought lacking in Schubert, stating broadly, “A person who is-confined for an offense prior to the imposition ’of sentence for said offense is entitled to credit against the term of his or her sentence for the entire period of such confinement.” § 18-1.3-405 (emphases added). Similarly, with its 1988 and 2009 amendments, the legislature identified precise scenarios under which courts must avoid awarding duplicative credit: when (1) an offender is already serv*206ing a sentence or is on parole or (2) an offender’s conduct has triggered parole revocation proceedings. See id.
¶86 I see nothing instructing courts to award credit against the sentences for multiple offenses charged in the confining jurisdiction but not toward those charged in other jurisdictions. To the contrary, the legislature, has required courts to award credit for confinement “for an offense” so long as the offender was confined “prior to the imposition, of sentence for [that] offense.” Id. And this makes sense: A defendant facing multiple charges is just as much confined “for an offense” whether the other charge preventing her release originated within the confining jurisdiction or without. Further, the inclusion of the two specific limitations to this general rale suggests legislative intent that there be no others, such as a geographic limitation. See Cain v. People, 2014 CO 49, ¶ 13, 327 P.3d 249, 253 (“[W]e read the General Assembly’s inclusion of a single, specific, narrow exception to mean that .the General Assembly intended that there be no other exceptions to the rule.... ”).
¶87 Thus, when multiple offénses' concurrently cause an offender’s confinement and only one of them results in a sentence, I conclude she is equally entitled to credit for her entire period of confinement concurrently caused "by those offenses, irrespective of where her confinement occurred. To effectuate the statutory purpose, she must receive credit somewhere. See § 18-1.3-405.
¶88 And while I must presume the General Assembly legislates with awareness of our case law, Vigil v. Franklin, 103 P.3d 322, 327 (Colo. 2004), three circumstances -suggest that the legislature’s amendments do not include or incorporate Massey’s geographic restriction. First, the current statute’s plain language covers the same territory as our cases with greater specificity—the 1986, 1988, and 2009 amendments correspond precisely to the concerns we raised in Schubert, 698 P.2d at 795, and Torand, 698 P.2d at 800-01. See Robles v. People, 811 P.2d 804, 806 (Colo. 1991) (holding that when the legislature adds more specific sections to a general section, it may indicate an intent to clarify the existing statute). Second, although we sometimes infer intent to ratify case law when the legislature amends a statute but leaves a previously interpreted provision unchanged, see Specialty Rests. Corp., 231 P.3d at 403, that inference is unjustified here because the provision Massey interpreted did change. Third, the current statute directly contravenes Massey by requiring some credit across jurisdictions: It explicitly contemplates that offenders serving an existing sentence in one jurisdiction will receive credit on that sentence for presentence confinement in another jurisdiction on a new charge.
¶89 Finally, I note that while I have now arrived at the position Justice Lohr advanced in his dissents to Massey and Freeman, I have done so because that is the outcome the PSOC statute’s current plain language best supports. The General Assembly has stated the required causal connection and has limited its effects where necessary. This court should go no further. That those limitations require results at odds with those the Massey test would have produced only further convinces me that the legislature did not intend for us to retain Massey’s geographic limitation.
¶90 Thus, I would conclude that section 18-1.3-405 does not require an offender seeking PSCG to. show she would have remained confined in the same jurisdiction on the charge for which credit is sought in the absence of any other charge. Instead, she must show that she was confined for an offense before being sentenced for that offense and that the offense shares a substantial nexus with the period of confinement for which she seeks credit.
D, Application
¶91 Applying my understanding of the substantial nexus test to the facts here, I conclude Torrez is entitled ■ to PSCC for the more than four years she spent confined before being sentenced on any offense—from her arrest in Denver on March 23, 2004, until she was sentenced in Jefferson County on July 25, 2008. The record demonstrates that throughout that period,' the offenses in both counties concurrently caused her confinement and that her confinement shared a *207substantial nexus with the Jefferson County offense. •
¶92 First, the parties do not dispute that Torrez was arrested on both warrants before being jailed in Denver. At all times before the Denver jury issued its not-guilty-by-reason-of-insanity verdict, Torrez was confined—wherever she may have found herself at the time—on both the Denver and Jefferson County charges. And following that verdict, the Jefferson County order continued to confine Torrez in CMHIP, where she spent nearly two more years before pleading guilty and receiving her sentence in Jefferson County. The Jefferson County offense was therefore an actual cause of her confinement. Second, no other offense had already initiated Torrez’s confinement, and until its resolution more than four years after her initial arrest, the Jefferson County offense continued to prevent her release. Thus, the Jefferson County offense satisfies the substantial nexus requirement in this context.
¶93 Moreover, because I agree with the court of appeals, People v. Torrez, 2012 COA 51, ¶ 17, 405 P.3d 243, and Justice Márquez, dis. op. ¶ 54 n.3, that time spent in CMHIP constitutes confinement under section 18-1.3-405 (the term’s ordinary meaning confirms that conclusion, and this court has never held otherwise), I believe Torrez -must receive credit for her period of confinement there as well. On these facts, so long as proceedings on the Jefferson County offense remained unresolved, Torrez was “confined for an offense prior to the imposition of sentence for said offense,” irrespective of where that confinement took place. Section 18-1.3-405 therefore entitles her to credit against the Jefferson County sentence for her entire period of confinement.
II. Conclusion
¶94 The majority today directs courts to award PSCC only when a defendant “would have been released ... had [the sentencing charges] not existed.” Maj. op. ¶ 3. This court has never employed such a test, and the statute does not compel it. Instead, I conclude the General Assembly’s amendments -to that statute demonstrate its intent to offer a broad grant of credit, which in turn entitles Torrez to PSCC on these facts. Therefore, I respectfully dissent. , .
I am authorized to state that JUSTICE GABRIEL joins in this dissent.
. The 1979 statute read as follows:
A person who is confined prior to the imposition of sentence is entitled to credit against the term of his sentence for the entire period of such confinement.' At the time of sentencing, the court shall make a finding of the amount of presentence confinement to which the offender is entitled and shall include such finding in the mittimus. Such period of confinement shall be deducted from the sentence by the department of corrections.
Ch. 157, sec. 7, § 16-11-306, 1979 Colo. Sess. Laws 664, 665-66.