SOSMAN, J. (concurring in part).

The present case calls on this court to craft working definitions of two terms used, but conspicuously not defined, in Crawford v. Washington, 541 U.S. 36, 53 n. 4, 68 n. 10 (2004) ( Crawford). In Crawford, the United States Supreme Court held that the introduction of "testimonial" statements made by a declarant who does not testify at trial is prohibited by the confrontation clause, unless the declarant is unavailable at trial and the defendant had a previous opportunity to cross-examine the declarant. Id. at 68. However, while providing various possible "formulations" of what statements might be considered "testimonial," the Court did not define the term or even signal any preference amongst the "formulations" offered. Id. at 51-52, 68. We do know that, whatever else may be included within the term "testimonial," it encompasses statements procured by way of police "interrogation," but, again, the Court expressly declined to define what it meant by police "interrogation" for these purposes. Id. at 52, 53 n. 4.

Although the exercise is fraught with uncertainty due to the Supreme Court's refusal to define the terms used in Crawford, courts across the country have had to interpret and apply the requirements of Crawford to the common fact pattern now before us: spontaneous utterances made by victims (and other eyewitnesses) to officers arriving at the scene in response to an emergency 911 call. Few courts have given Crawford's undefined terms the extreme interpretation adopted today. While today's opinion properly allows the parties to reopen the record below to present additional evidence in light of Crawford, it erroneously signals that, unless the police questions were specifically aimed at defusing a "volatile" scene or attending to the complainant's medical needs, the police questions automatically amounted to "interrogation" and the statements made in response were therefore "testimonial." In accord with many other courts, I believe that "police interrogation" does not encompass the basic, immediate, on-scene questioning of persons present in an attempt to get the gist of what is happening or has just happened, i.e., to ascertain why police were called to the scene and what steps need to be taken in response.

And, if the declarant's statements are not the product of "police interrogation," and we must therefore turn to the more general "formulations" of the term "testimonial," today's decision adopts a definition of "testimonial" that goes well beyond any of those formulations. It also ignores the fact that the predicates for introduction of a statement as a spontaneous utterance will, as a practical matter, make it unlikely in the extreme that the statement was "testimonial," even if the statement was made to the police. This, too, is the position now being adopted by many courts across the country. In accord with those other jurisdictions, I do not interpret Crawford to prohibit the introduction of spontaneous utterances given in response to this kind of preliminary police questioning, and would, on the record compiled to date, allow the introduction of at least the initial statements made to the police by the complainant. I thus agree that the case needs to be remanded for a more complete record addressing the requirements of Crawford, but do not agree with today's articulation of those requirements.

1. " Police interrogation." Statements procured by "police interrogation" are one subcategory of "testimonial" statements. Crawford, supra at 68. While recognizing that there could be "various definitions of 'interrogation'" in this context, the Court declined to suggest — let alone adopt — any definition. Id. at 53 n. 4. The only guidance provided tells us that we must use the term "in its colloquial, rather than any technical legal, sense." Id. Lacking any definition, we must consider the Court's reasons for treating statements procured by "police interrogation" as inherently "testimonial," and, with particular reference to the issue before us, how those reasons are or are not implicated in the kind of preliminary questioning conducted by police on arrival at the scene in response to an emergency 911 telephone call.

The Court included responses to "[p]olice interrogations" within the category of "testimonial" statements because such interrogations "bear a striking resemblance to examinations by justices of the peace in England," noting that justices of the peace performed investigative functions in that earlier era comparable to investigative functions now performed by police. Crawford, supra at 52-53. While justice of the peace examinations were investigative in the sense that they were oriented toward establishing the case for the prosecution and not conducted in the manner of neutral magistrates, they do not appear to have included on-scene preliminary questioning in the immediate aftermath of an unsolved crime, but were instead conducted only after someone had been accused and brought before them. See 4 W. Blackstone, Commentaries 293-296 (8th ed. 1878); 2 M. Hale, Pleas of the Crown 120-121 (1800). And as described in Crawford, such examinations were frequently under oath, and reduced to writing to be "read" in court. Crawford, supra at 43-44. See Anderson v. State, 111 P.3d 350, 357-358 (Alaska Ct.App. 2005) (Mannheimer, J., concurring) ("police interrogation" in Crawford refers to "the kind of formal, systematic questioning that was characteristic of the English inquisitorial practices that prompted the enactment of the [c]onfrontation [c]lause"). The "police interrogation" that resembles a justice of the peace examination is thus the more thorough, complete interview, often conducted later at the station and reduced to writing or recorded.

The Court's reasons for applying the confrontation clause to statements procured by "police interrogation" also included a more general concern about the evil the confrontation clause was intended to address. "Involvement of government officers in the production of testimony with an eye toward trial presents unique potential for prosecutorial abuse — a fact borne out time and again throughout a history with which the Framers were keenly familiar." Crawford, supra at 56 n. 7. However, when the police arrive at the scene and ask basic, preliminary questions (e.g., "What's going on?"), they do not know whether they will be told about a crime, an accident, a misunderstanding or a false alarm; nor do they know whether anyone will report that there is some injury or imminent danger that needs their attention. When the police are trying to ascertain such basic information, they do not yet have "an eye toward trial," id. — they are trying to find out, among other things, whether there is even any crime to be investigated, and they are simultaneously trying to ascertain whether there is any ongoing threat or need for medical attention. Such preliminary questions on arrival at the scene ("What's happening?" "What did he do?" "Where is he now?") are not intended to produce "testimony" but to assess the situation and inform the officers' decision as to what immediate steps they must take. I see no "unique potential for prosecutorial abuse," id., in such preliminary inquiries.

Once the police are in fact investigating an alleged crime and preparing a case for potential prosecution, police questioning that takes on some aura of formality, solemnity, or (at a minimum) thoroughness, is what makes it "interrogation" for purposes of the confrontation clause. As today's decision acknowledges, the dictionary definition of "interrogation" is "question[ing] typically with formality, command, and thoroughness for full information and circumstantial detail." Ante at 7-8, quoting Webster's Third New Int'l Dictionary 1182 (1993). See American Heritage Dictionary 685 (4th ed. 2000) (defining "interrogate" as "[t]o examine by questioning formally or officially"); Merriam-Webster's Dictionary of Law (1996) ("to question formally and systematically"). That is the "colloquial" understanding of what "interrogation" entails, Crawford, supra at 53 n. 4, and that is the definition we should be applying to comply with Crawford.

The proposition that "police interrogation" involves structured, thorough, and detailed questioning has considerable support in cases to date. See, e.g., Mungo v. Duncan, 393 F.3d 327, 336 n. 9 (2d Cir. 2004), cert. denied sub nom. Mungo v. Greene, 544 U.S. 1002 (2005) (rejecting proposition that "interrogation" encompasses "any asking of questions," in favor of dictionary definition "to question typically with formality, command, and thoroughness for full information and circumstantial detail," as that definition was "more consistent" with types of testimonial statements mentioned in Crawford); Anderson v. State, supra at 353 ("interrogation" refers to "category of formal, official, and systematic questioning"); People v. Corella, 122 Cal. App. 4th 461, 468 (2004) ("police interrogation" under Crawford "requires a relatively formal investigation where a trial is contemplated"); United States vs. Webb, D.C. Sup. Ct. No. DV-339-0400 (Nov. 9, 2004); State v. Hembertt, 269 Neb. 840, 850 (2005) ("structured police questioning"). See also Hammon v. State, 829 N.E.2d 444, 457 (Ind. 2005) ("police interrogation" for purposes of confrontation clause "is properly limited to attempts by police to pin down and preserve statements").

Some courts have taken the extreme position that essentially all police questioning of witnesses constitutes "police interrogation." See, e.g., Moody v. State, 277 Ga. 676, 679-680 n. 6 (2004) ("police interrogation" encompasses "field investigation of witnesses"); State v. Sutton, 609 S.E.2d 270, 275 (N.C.Ct.App. 2005) (when police question crime victim, "that officer clearly has 'an eye toward trial'"); Wall v. State, 143 S.W.3d 846, 851 (Tex.Ct.App. 2004). If the Supreme Court intended to sweep any and all police questioning within the ambit of "interrogation" and render all responses thereto "testimonial," that would have been an easy definition to craft. The Court did not do so. Instead, the Court indicated that the term police "interrogation" was to be used "in its colloquial, rather than any technical legal, sense," Crawford, supra at 53 n. 4, and we must therefore focus on whether a lay person would consider a particular form of police questioning to be "interrogation."

Today's decision takes the position that "police interrogation" encompasses "all law enforcement questioning related to the investigation or prosecution of a crime," including "preliminary fact gathering and assessment whether a crime has taken place," ante at 8, 9, an approach that is essentially as extreme as the above-cited cases. From this admittedly "expansive" definition, ante at 8, the court today does carve out an exception for circumstances where, in addition to investigating criminal conduct, the police are simultaneously addressing a "volatile scene" or attending to a complainant's medical needs, because those tasks come within "the government's peacekeeping or community caretaking function." Ante at 9. Of course, when criminal conduct has engendered some imminent threat to safety, ante at 9 n. 4, or has resulted in some form of medical emergency, police who respond to the scene are not asking questions that are unique to their caretaking as opposed to their investigative function. They are performing both functions at the same time, and customarily using the same questions. It strikes me as arbitrary, unsupported by any language in Crawford, and unrelated to the purposes served by the confrontation clause, to have admissibility of a witness's response to such questions hinge on the fortuity that the criminal activity that the police have come to investigate happens simultaneously to involve some aspect of their caretaking function.

In the wake of today's opinion, we will be embroiled in arguments over whether particular questions at the scene had some possible link to the community caretaking function, despite their obvious overlap with criminal investigation. By their nature, for example, questions about what was done to the victim during the attack (how many blows, to what portion of the victim's body, with what weapon) simultaneously seek information about what injuries the victim sustained that might need medical attention and information that provides the basic elements of the crime committed. I foresee much hairsplitting, as attorneys wrangle and judges attempt to separate the "caretaking function" questions from the "investigative" questions. Nothing in Crawford suggests the parsing exercise that today's opinion has launched, and it is an exercise that will prove difficult to perform.

Moreover, today's distinction (between questions asked to assess "whether a crime has taken place" and questions asked to address a public safety threat or a medical emergency at the scene of that crime) makes the term "police interrogation" depend not on the questions that the police are asking but on the responses they get to those questions. For example, if the police respond to a call that a woman has been attacked in her home by a knife-wielding intruder, they will presumably ask the victim where the alleged perpetrator is — if she tells the police that the armed perpetrator is still hiding somewhere in the house, she has informed the police of a potentially "volatile scene" (thus bringing the exchange within some "caretaking" function), but if she tells the police that the perpetrator fled on foot in a particular direction, she has given them information pertinent only to their investigative function. Similarly, if the police ask that victim what the perpetrator did, the victim's response that he slashed her arm while she tried to fend him off will give the police information about a wound that may need medical care, but a response that he pointed the knife at her will merely describe an assault that did not lead to any physical injury requiring medical intervention. "Interrogation" surely refers to the nature of the questioning conducted by the police, and is not dependent on the substantive contents of the witness's answers to police questions.

Questioning by police on their arrival at the scene simultaneously addresses a range of issues that might lead to further police action of various forms. When the police ask those questions, they do not yet know what answers they will get or what actions they will take. When responding to an emergency call, they are addressing that emergency, not getting ready for trial. Police questioning in these circumstances is quick, open ended, impromptu, informal, not under oath, not reviewed by the witness, not detailed or thorough. Lay persons would not call this "interrogation," and neither should we. It bears no resemblance whatsoever to justice of the peace examinations; it is not the historical evil that the confrontation clause was meant to cure; and it is not what Crawford meant to address.

Many courts have treated such preliminary questioning at the scene as outside the purview of "police interrogation." See, e.g., Mungo v. Duncan, 393 F.3d 327, 336 n. 9 (2d Cir. 2004), cert. denied sub nom. Mungo v. Greene, 544 U.S. 1002 (2005) (initial questioning at scene, during pursuit, and immediately after apprehension of suspects not "interrogation," but victim's statements probably became "testimonial" when officer pressed victim for additional clarification of his ambiguous identification, as that line of questioning entailed "greater formality with a view to creating a record and proving charges"); Anderson v. State, 111 P.3d 350, 356 (Alaska Ct.App. 2005) ("excited responses to brief on-the-scene questioning by police officers" not "testimonial"); People v. Corella, 122 Cal. App. 4th 461, 468 (2004) (preliminary questions at scene shortly after crime is "an unstructured interaction," not "interrogation"); United States vs. Webb, D.C. Sup. Ct. No. DV-339-0400 (Nov. 9, 2004) ("it is exceedingly unlikely that the Supreme Court intended to exclude from evidence excited utterances made during investigatory questioning at the scene of a crime soon after the criminal event," finding nontestimonial victim's response to officer's general questions of "what happened" and "why"); Hammon v. State, 829 N.E.2d 444, 457 (Ind. 2005) ("responses to initial inquiries by officers arriving at a scene are typically not testimonial"); State v. Hembertt, 269 Neb. 840, 853-854 (2005) (victim's statement to police on arrival not "testimonial," but subsequent interview of victim after suspect apprehended properly excluded as "testimonial"); Key v. State, 173 S.W.3d 72, 75 (Tex.Ct.App. 2005) ("initial police-victim interaction at the scene of an incident is not an interrogation"). See also People v. Mackey, 5 Misc. 3d 709, 712 (N.Y. Crim. Ct. 2004) ("responses to police officers during a preliminary field investigation" not testimonial if encounter "lack[s] the requisite formality to constitute a police interrogation"); State v. Forrest, 164 N.C. App. 272, 280 (2004), aff'd, 611 N.C. 833 (2005) (where victim initiated spontaneous statements to police when they arrived at scene, follow-up questions and answers were not testimonial). Indeed, this interpretation is now recognized as "the emerging majority view" of Crawford as it relates to initial police questioning at the scene of a crime. Anderson v. State, supra at 356. See Spencer v. State, 162 S.W.2d 877, 881 (Tex.Ct.App. 2005).

It may take very few questions and very little time before preliminary police questioning ripens into something more thorough, more pointed, or more detailed, and thus crosses the line into "interrogation." See, e.g., Mungo v. Duncan, supra at 336 n. 9; State v. Hembertt, supra at 850. The questioning need not be as structured, or the setting as formal, as in Crawford itself. At some point after the police have the rudimentary information gleaned on their arrival at the scene of a crime, it is customary that the police seek to obtain some more detailed, orderly, and complete account of the incident, asking specific and probing questions to address any gaps or inconsistencies in the witnesses' answers, and the police then commonly preserve the witnesses' information in a subsequent report. In its "colloquial" sense, that kind of questioning is "interrogation." Crawford, supra at 53 n. 4. However, the "colloquial" understanding of the term "police interrogation" does not include the officer, responding to a report of some domestic disturbance, who enters the home and addresses all persons present with the question, "What's going on?"

While it may in some cases be difficult to draw a precise line as to when preliminary questioning ends and "interrogation" begins, it is a line that we can discern, and to the extent that it is difficult to do so, it is at least a distinction that has some relevance to the underlying purpose of the confrontation clause. Where "interrogation" does not and should not encompass any and all police questioning, a line must be drawn somewhere, and I believe that other courts distinguishing between preliminary on-scene questioning and "interrogation" are focusing on an appropriate distinction. Today's proposed line — determining whether a particular question can be justified by reference to community caretaking even if it simultaneously overlaps with criminal investigation — will be even more difficult to draw, and the distinction is irrelevant to the purposes underlying the confrontation clause.

The record in the present case was sparse with respect to how detailed or thorough the police questioning of the complainant ultimately became. The evidence concerning her statements to the responding officers was developed for the purpose of determining whether the requisites for the spontaneous utterance exception had been established, not with an eye toward whether (or at what point) the questioning that elicited those spontaneous utterances had become a "police interrogation." In light of Crawford, I agree that the record needs to be reopened to address that issue. However, I do not think that resolution of the issue hinges on such things as how the ambulance came to be summoned, or whether the police questions were motivated by concern for the complainant's medical needs. Rather, what should be examined is whether and at what point the questions being asked of the complainant became sufficiently detailed, structured, or pointed that a layperson would describe her as being "interrogated" by the police. Once the police questioning crossed that line, the complainant's responses were, by definition, "testimonial" under Crawford.

2. " Testimonial." If it is determined that the complainant made some statements prior to the onset of any "police interrogation," the judge must still assess whether those statements are nevertheless "testimonial" within the meaning of Crawford. Responses to "police interrogation" comprise merely one subset of "testimonial" statements, and a statement made to police that is not the product of "police interrogation" may still be "testimonial" for some other reason.

Again operating without any definition of "testimonial," we must start the analysis by reviewing the three differing general formulations of "testimonial" noted by the Court in Crawford. Although not expressly approved or adopted by the Court, those formulations were at least given some general imprimatur by way of quotation within the text of the opinion. They are as follows: "[1] ex parte in-court testimony or its functional equivalent — that is, material such as affidavits, custodial examinations, prior testimony that the defendant was unable to cross-examine, or similar pretrial statements that declarants would reasonably expect to be used prosecutorially"; [2] "extrajudicial statements . . . contained in formalized testimonial materials, such as affidavits, depositions, prior testimony, or confessions"; [3] "statements that were made under circumstances which would lead an objective witness reasonably to believe that the statement would be available for use at a later trial" (citations omitted). Crawford, supra at 51-52.

Today's decision reaches beyond these formulations and declares a statement "testimonial" whenever "a reasonable person in the declarant's position would anticipate the statement's being used against the accused in investigating and prosecuting a crime." Ante at 12-13. This articulation of the test is drawn from United States v. Cromer, 389 F.3d 662, 675 (6th Cir. 2004) ( Cromer). Under this test, essentially anything said to law enforcement or government officials that pertains to any form of criminal activity is "testimonial" — after all, a reasonable person would anticipate that giving any government official information concerning criminal conduct will engender at least an investigation and, if the investigation bears fruit, a prosecution. Under Cromer's definition, any "statement made knowingly to the authorities that describes criminal activity is almost always testimonial." Id., quoting Friedman, Confrontation: The Search for Basic Principles, 86 Geo. L.J. 1011, 1042 (1998). Indeed, in a more recent case, the United States Court of Appeals for the Sixth Circuit confirmed just how expansive its definition of "testimonial" was. In considering whether statements made during an emergency 911 call and two later statements made in person to the police were testimonial, the court noted that "[o]n all three occasions, [the declarant] made the statements to government officials: the police. This fact alone indicates that the statements were testimonial" (footnote omitted). United States v. Arnold, 410 F.3d 895, 903 (6th Cir. 2005).

Before adopting this all-encompassing definition, we should first compare it to the formulations articulated in Crawford itself. The closest kin to the Cromer articulation, but with a distinct and important difference, is the third Crawford formulation: statements "made under circumstances which would lead an objective witness reasonably to believe that the statement would be available for use at a later trial." Crawford, supra at 52. While this Crawford formulation focuses on the declarant's appreciation of the potential use of the statement "at a later trial," id., Cromer sweeps more broadly and renders testimonial any statement made by a declarant who should reasonably anticipate use of the statement "in investigating and prosecuting the crime." Cromer, supra at 675. Any form of accusation or statement concerning criminal activity would cause a reasonable person to anticipate that an investigation will follow, but not every such accusation or statement is made with an eye toward trial. Indeed, statements made in the presence of police arriving while the crime is still in progress (e.g., a hostage screaming, "Watch out! They both have guns! They threatened to shoot me if anyone comes near!") would be testimonial under Cromer — the declarant would reasonably expect that the police would later use that information to investigate and ultimately prosecute the perpetrators. The Cromer approach thus treats as "testimonial" a vast array of informal statements that are not made for purposes of trial, whereas Crawford's third formulation of "testimonial" (and the historical evil to which the confrontation clause is addressed) is focused on statements made and preserved specifically for use at trial.

Moreover, whatever support for the Cromer definition can be found in the third Crawford formulation, Cromer is plainly at odds with the other Crawford formulations. The second such formulation expressly notes formality as a hallmark of "testimonial" statements: statements "contained in formalized testimonial materials, such as affidavits, depositions, prior testimony, or confessions" qualify as "testimonial" (emphasis added). Crawford, supra at 52, quoting White v. Illinois, 502 U.S. 346, 365 (1992) (Thomas, J., concurring in part and concurring in judgment). In a similar vein, the Court cited a dictionary definition of "testimony" ("[a] solemn declaration or affirmation made for the purpose of establishing or proving some fact"), and noted that "[a]n accuser who makes a formal statement to government officers bears testimony in a sense that a person who makes a casual remark to an acquaintance does not" (emphasis added). Crawford, supra at 51.

Justice Thomas's concurring opinion in the White case (in which Justice Scalia, the author of Crawford, joined) stressed that the confrontation clause was aimed "only" at the "discrete category" of "formalized testimonial materials, such as affidavits, depositions, prior testimony, or confessions," and that it should "not be construed to extend beyond the historical evil to which it was directed." White v. Illinois, supra (Thomas, J., concurring in part and concurring in judgment). The recognition of that same "formulation" in Crawford suggests a recognition of that same limitation — the "historical evil" addressed by the confrontation clause was the use of ex parte testimony taken by justices of the peace, and only "formalized testimonial materials" that resemble, in some fashion, those justice of the peace examinations are the subject of the confrontation clause prohibition. If we are to give any weight at all to the second Crawford formulation, we cannot adopt Cromer's all-encompassing definition of "testimonial," as it completely disregards any element of formality, solemnity, or the like.

The first Crawford formulation does not support Cromer either: "ex parte in-court testimony or its functional equivalent — that is, material such as affidavits, custodial examinations, prior testimony that the defendant was unable to cross-examine, or similar pretrial statements that declarants would reasonably expect to be used prosecutorially." Crawford, supra at 51. While this formulation refers to a declarant's anticipation that the statement would be "used prosecutorially," a term that would arguably encompass use of the statement prior to actual trial and perhaps as early as the mere investigative phase, its actual list of examples again consists of formalized statements — in-court testimony, affidavits, and custodial examinations. The broader reference to the statements that declarants would expect to be "used prosecutorially" is expressly linked to that prior list: "or similar pretrial statements that declarants would reasonably expect to be used prosecutorially" (emphasis added). Id. In other words, not every statement that one might expect to be "used prosecutorially" is "testimonial," but only those statements that are "similar" to in-court testimony, affidavits, or custodial examinations. Thus, the first Crawford formulation also has an emphasis on some degree of formality or some obvious connection with court proceedings. It does not support Cromer's far-reaching definition of "testimonial."

Whereas Cromer ignores Crawford's first two formulations and then notably expands on Crawford's third formulation, we should strive for a definition that harmonizes all three formulations. The third formulation focuses on whether the declarant would reasonably appreciate a likelihood that the statement would be used "at a later trial," and the other two formulations signal that the formal, solemn, or official nature of the statement is what would ordinarily alert a reasonable person to that likelihood. See United States v. Saget, 377 F.3d 223, 228-229 (2d Cir. 2004), cert. denied, 543 U.S. 1079 (2005) (determinative factor is "declarant's awareness or expectation that his or her statements may later be used at trial"). See also Hammon v. State, 829 N.E.2d 444, 456 (Ind. 2005) ("'testimonial' statement is one given or taken in significant part for purposes of preserving it for potential future use in legal proceedings"). That a precise definition is elusive — so elusive that the Court declined to give one, despite the obvious adverse consequences of failing to do so — does not allow us to expand on the Crawford formulations. The Cromer definition represents a significant and unwarranted extrapolation from Crawford, and I see no reason to stray beyond the guideposts that Crawford itself has provided.

The ramifications of Cromer also place it squarely at odds with other courts addressing statements made during emergency calls to police for help. Applying its own Cromer definition in a literal manner, the United States Court of Appeals for the Sixth Circuit recently pronounced that statements made during a 911 call, notwithstanding that their purpose "may have been to secure assistance," are inherently "testimonial" because they were made to "the police." United States v. Arnold, 410 F.3d 895, 903, 904 (6th Cir. 2005). When addressing similar communications seeking emergency assistance, other courts have found them not to be "testimonial" where the declarant was focused on obtaining urgently needed assistance, not contemplating the statement's potential use at a later trial. See, e.g., Leavitt v. Arave, 383 F.3d 809, 830 n. 22 (9th Cir. 2004); People v. Corella, 122 Cal. App. 4th 461, 468 (2004); State v. Barnes, 854 A.2d 208, 211 (Me. 2004); People v. Moscat, 3 Misc. 3d 739, 745 (N.Y. Crim. Ct. 2004); State vs. Newall, Ohio Ct. App. No. 2004 CA 00265 (May 31, 2005). See also United States v. Arnold, supra at 914 (Sutton, J., dissenting) (most 911 calls not testimonial, as "a 911 call generally will be a plea for help, not an effort to establish a record for future prosecution").

Given the complex, fact-specific, nuanced approach taken by other courts struggling to interpret and apply Crawford, Cromer's simplistic everything-said-to-police-is-testimonial tack appears out of step. While no one can be sure how the Supreme Court will ultimately define "testimonial," I doubt that the Court will adopt a definition that goes beyond what Crawford actually says and that is so far at odds with the interpretation of so many courts. There is a spectrum of opinion as to what Crawford means, with Cromer standing at one extreme end of that spectrum. I am not prepared to endorse that extreme end, but would hew closely to the formulations articulated in Crawford itself.

3. Relationship between "testimonial" statements and spontaneous utterances. With the overruling of the "reliability" test of Ohio v. Roberts, 448 U.S. 56 (1980), the Crawford analysis does not hinge on the rationale justifying any particular hearsay exception — we must consider whether a statement is "testimonial," not whether there is a sound basis for exempting it from the rule against hearsay. However, with regard to the specific hearsay exception of spontaneous utterances, we should not ignore the fact that the prerequisites for that exception are wholly incompatible with the features that make a statement "testimonial." To qualify as a spontaneous utterance, the statement must be made when the declarant is still under the influence of an exciting event, before having time to contrive or fabricate the statement, and it must tend to qualify, characterize and explain the exciting event. See Commonwealth v. King, 436 Mass. 252, 254-255 (2002), and cases cited. Ordinarily, these prerequisites are met by demonstrating that, in the immediate aftermath of some traumatic event, the declarant was in a highly distraught state, blurting out statements concerning that event without opportunity or capacity for reflection.

By their nature, spontaneous utterances lack the formality or solemnity that has been associated with "testimonial." They also fail the test of what the declarant would "reasonably expect" in terms of the use of the statement at a later trial. When the declarant's emotional state is such that it would satisfy the spontaneous utterance exception, the declarant does not (and reasonably would not) entertain any expectations about whether his or her statement will be used at a possible future trial — the whole premise of spontaneous utterances is that the witness is not capable of such reasoning at the time he or she is speaking.

Based on the lack of formality and the absence of any expectation on the part of the declarant that his spontaneous utterances will be used at trial, various courts have expressed the view that spontaneous utterances are not testimonial or are highly unlikely to be testimonial. See, e.g., United States v. Brown, 322 F. Supp. 2d 101, 105 n. 4 (D. Mass. 2004) (in dicta, "doubtful" that Crawford would apply to spontaneous utterances); Anderson v. State, 111 P.3d 350, 354 (Alaska Ct.App. 2005) (finding that statement was excited utterance makes it "inconsistent" with Crawford formulation of "testimonial"); State v. Aguilar, 210 Ariz. 51, 53 (2005) (excited utterance not "even remotely similar to most of what Crawford offers as an example of a testimonial statement"); People v. Corella, supra at 469 ("difficult to identify any circumstances under which a . . . spontaneous utterance would be 'testimonial,'" as spontaneous utterances are "not made in contemplation of their 'testimonial' use in a future trial"); Hammon v. State, 829 N.E.2d 444, 453 (Ind. 2005) ("declarant of an excited utterance will ordinarily lack the requisite motive" to make testimonial statement); Commonwealth v. Eichele, 66 Pa. D. C. 4th 460, 469 (2004) ("excited utterance is at the opposite end of the hearsay spectrum from testimonial hearsay"); State vs. Anderson, Tenn. Crim. App. No. E2004-00694-CCA-R3-CD (Jan. 27, 2005) (essential characteristics of excited utterances "conflict with the characteristics that would make them testimonial" and "cannot be testimonial in that such a statement has not been made in contemplation of its use in a future trial").

Notwithstanding that the prerequisites for admission as a spontaneous utterance are utterly disconnected from the formulations of "testimonial," there are still circumstances where a spontaneous utterance must be classified as "testimonial." For example, if the nature of the police questioning is such that it qualifies as "police interrogation," the fact that the witness is still under the sway of a recent traumatic event will not operate to take the witness's answers outside the bounds of "testimonial." See Hammon v. State, supra at 455-456; State v. Hembertt, 269 Neb. 840, 850-851 (2005). There are times where, notwithstanding a witness's emotional condition, the police attempt a systematic questioning of the witness, pressing for specifics and detail, and thereby conduct a "police interrogation," to which the witness responds with spontaneous utterances. I thus do not accept the Commonwealth's argument that all spontaneous utterances are automatically beyond the purview of Crawford. See Spencer v. State, 162 S.W.3d 877, 881 (Tex.Ct.App. 2005) (rejecting "bright-line rule that excited utterances can never be testimonial," but recognizing nature of spontaneous utterance as factor to be considered in assessing whether statement "testimonial"). However, if the declarant of a spontaneous utterance is not being subjected to police interrogation, it is unlikely in the extreme that the circumstances will satisfy the more general formulations of "testimonial." It is still a separate inquiry, and the defendant is entitled to have the Crawford analysis conducted independently of the evidentiary analysis under the spontaneous utterance exception, but it is the rare circumstance — falling almost into the category of "never say never" — that a spontaneous utterance made outside of any "police interrogation" will qualify as "testimonial." See United States v. Arnold, 410 F.3d 895, 914 (6th Cir. 2005) (Sutton, J., dissenting) (number of "testimonial" statements that also qualify as spontaneous utterances "may be something approaching a null set").

Again, I agree that the present matter needs to be remanded so that that separate Crawford analysis can be made on a record addressing the issue, but I see nothing in the record to date to suggest that this particular declarant had any "awareness or expectation that [her] statements [would] later be used at a trial," United States v. Saget, 377 F.3d 223, 228 (2d Cir. 2004), cert. denied, 543 U.S. 1079 (2005), or that her purpose in making them was to preserve her testimony for use in future legal proceedings, Hammon v. State, supra at 457-458. Not only was her emotional state such as to qualify for the spontaneous utterance exception, but she had not summoned the police or expected their arrival, and there is nothing to indicate that she even intended to press charges at the time.

4. Ramifications of overly broad definitions of "testimonial." Today's decision, with its expansive interpretation of both "police interrogation" and the broader concept of "testimonial," recognizes that its approach will have a far-reaching impact on the prosecution of certain types of crimes, in particular crimes involving domestic violence. Ante at 14. Although domestic violence prosecutions are the most obvious category of cases in which a victim's refusal to testify often forces the prosecution to rely heavily on spontaneous utterances to police at the scene, it is not the only category of cases affected by today's decision. Witnesses are reluctant to testify in a variety of circumstances, and may often claim a valid privilege at trial, or simply absent themselves and thereby become unavailable. Where those same witnesses have made spontaneous utterances to police at the scene, introduction of those spontaneous utterances may be essential to a successful prosecution of the case. The need to comply with Crawford is, of course, paramount, but any unnecessarily expansive interpretation of Crawford will hamper the ability of the criminal justice system to address crimes involving domestic violence or gang violence, and to prosecute crime in the face of successful witness intimidation.

While recognizing this problem, today's opinion pronounces that the criminal justice system must simply "adjust." Ante at 14. We must, of course, "adjust" to whatever new requirements Crawford has in fact laid down, but we need not impose restrictions any more onerous than what the Supreme Court intended. In the absence of any definitions in Crawford itself, we are caught between the Scylla of violating a defendant's confrontation rights if we interpret "testimonial" and "police interrogation" too narrowly and the Charybdis of needlessly thwarting meritorious prosecutions if we exclude evidence based on too expansive a definition of those same terms. Pending further clarification from the Supreme Court, we must attempt to steer the precise middle course between the two. Today's opinion does not strike me as that middle course, but as an overly expansive reading of Crawford that will unnecessarily exclude critical evidence from a significant array of cases. Courts around the country, faced with the same dilemma we are, have developed a more moderate approach in applying Crawford's undefined terms to the very common scenario posed by the present case. Today's decision eschews moderation, instead adopting a definition of "testimonial" that stands at the extreme edge of current jurisprudence and that goes beyond the actual language and reasoning of Crawford. As illustrated by the ever-increasing number of cases addressing confrontation clause issues in the wake of Crawford, this is not an academic debate, but an issue of tremendous pragmatic consequence to the trial of criminal cases. Given those consequences, I would not undertake any creative extrapolation from Crawford. I thus concur with the result in today's opinion — a remand for development of a record that will address the confrontation clause issue in accordance with Crawford — but do not join in the court's assessment of what that Crawford analysis entails.