The International IJSLL (print) issn 1748-8885

IJSLL (online) issn 1748-8893

Journal of
Speech, Article
Language
and the Law

‘You have the right to remain silent…’

but only if you ask for it just so:
the role of linguistic ideology in American police
interrogation law

Janet Ainsworth

Abstract
American constitutional law provides protection for arrested persons from coercive
police interrogation by giving them the right to refuse to answer questions and to the
assistance of a lawyer during questioning. Once these Miranda rights are invoked
by the interrogated suspect, the police must cease questioning immediately. In recent
years, however, courts have adopted restrictive definitions of the kind of invocation
that will count as legally efficacious. In using hyper-literal parsings of the language
of suspects who attempt to assert their rights, these judges fail to utilize ordinary
norms of conversational implicature and ignore the power asymmetries inherent
in custodial police interrogation. Instead, these strained interpretations result in
part from an ideology of language that presumes that language is a transparent
medium of communication. Because direct and unmodified locutions are consequently
privileged, most attempts to claim constitutional rights are found to be defective. As
a result, the constitutional protections of Miranda are practically unattainable by
most arrestees.
keywords: police interrogation, miranda rights, ideology of language

Affiliation

Seattle University School of Law

email: jan@seattleu.edu

IJSLL vol 15.1 2008 1–21 doi : 10.1558/ijsll.v15i1.1

©2008, equinox publishing LONDON
2 The international journal of speech, language and the law

Legal theory conceptualizes language as a tool for the creation of and extinc-
tion of rights and obligations. Given this explicit link between the power of
law and its use of language as a vehicle for instantiating that power, linguistic
philosophy and analysis have had obvious implications in considering the
legal consequences of particular language usages. For example, J. L. Austin, in
his taxonomy of the performative uses of language, drew upon legal language
as quintessential exemplars of words that did things (Austin 1962). ‘I now
pronounce you man and wife’, or ‘We the jury find the defendant guilty’, are
classic examples of Austinian speech acts that, assuming appropriate felicity
conditions, bring specific legal consequences into being merely by their utter-
ance. By using this sort of speech act, speakers become juridical actors whose
speech is given operative significance by the law.
The use of a speech act to call a legal state of affairs into being is relatively
unproblematic as long as the speaker understands this practice as a verbal recipe
for achieving a particular sought-after legal result. The patterned recitation of
a particular utterance, constrained in its efficacy both by the social role of the
speaker and by the time, place, and circumstances of the utterance, gives would-
be legal actors the ability to manage legal relations. In the examples given, the
minister performing the marriage and the presiding juror announcing the
verdict are consciously aware of the fact that what they say in the context in
which they are saying it will produce – indeed, is explicitly intended to produce
– a specific, desired legal outcome. Legal language of this nature performs a
kind of ‘word magic’ (cf. Ogden and Richards 1989).
Not all speech acts with legal consequences occur within a ritualized,
scripted context, however. In cases other than these scripted situations, using
an unintentionally legally infelicitous utterance may cause speakers to fail to
achieve their desired legal outcomes and instead expose them to unintended
and unforeseen legal consequences. These unfortunate results are particularly
likely when the speaker lacks the legal training needed to know how to shape
legally felicitous speech acts. Legally naive speakers are often unaware that
a script exists prescribing the language needed for legal efficacy. When their
language fails to track the script, they are unable to achieve their desired
legal ends.
American criminal procedure governing police interrogation provides a good
example of a recurring legal situation in which legally naïve speakers with lim-
ited or no awareness of the script needed for a legally effective speech act must
nevertheless attempt to use language in such a way as to achieve their desired
legal result. People in police custody are protected during police interrogation
by the constitutional rules set out in Miranda v. Arizona, but the police are
unconstrained by those rules unless and until the arrestee says certain things
that call those protections into being. As a consequence, courts have needed to
 linguistic ideology in American police interrogation law 3

determine the contours of the ‘script’ necessary in this context and thus what an
arrestee must say in order successfully to exercise the constitutional Miranda
rights during police interrogation.
To understand the dimensions of this problem, it is necessary to map out the
constitutional framework in which the Miranda rights are embedded. The Fifth
Amendment of the United States Constitution guarantees that people cannot be
forced by the government or its agents to incriminate themselves. In addition,
under the Due Process Clause of the Fourteenth Amendment, confessions
are not considered legally valid and admissible into evidence unless they are
voluntary – that is, the product of the free will of the confessing party. Consider,
however, the reality of the situation for a suspect in police custody undergoing
interrogation. Under the inherently intimidating and coercive atmosphere of
incommunicado police interrogation, it is highly questionable that any resulting
confession could be considered the product of the unfettered free will of the
suspect. The Supreme Court was not blind to this reality as it struggled with
crafting procedural rules to govern police interrogation. Yet, had they followed
this insight to its logical conclusion, the Supreme Court would have had to
acknowledge that custodial interrogation is inconsistent with the procurement
of truly voluntary confessions.
Had the Supreme Court held that custodial interrogation inevitably taints the
voluntariness of a confession, the police would have had to forego such interro-
gation entirely. Obviously, the costs imposed on law enforcement investigation
rendered that an unpalatable option for the Court. As a practical matter, police
interrogation results in the acquisition of much valuable evidence that would
be otherwise unobtainable. The Miranda decision governing the conduct of
custodial police interrogation was intended as a compromise between the
extremes of outright banning police interrogation and allowing it to proceed
without any legal constraint at all. Custodial police interrogation of an arrested
person, often for hours on end, was fully acknowledged by the Miranda Court
to be so inherently coercive as to create a presumption that a resulting confes-
sion was involuntary unless the suspect was explicitly told prior to questioning
that he need not answer questions and that he had the right to consult with an
attorney for advice. In the Court’s view, preceding police questioning with this
information – the now-famous Miranda warnings – mitigated the oppressive
pressure to talk that might taint the voluntariness of any resulting confession.
Suspects needed to be specifically informed that they were under no obligation
to answer questions, but that if they did, their answers could be used to their
detriment in court. Additionally, they had to be told that they were entitled
to call upon an attorney to advise them as to the best course of action to take
during interrogation, and that appointed counsel would be provided if they had
no funds to hire legal counsel. The Court presumed that suspects armed with
4 The international journal of speech, language and the law

this information about their legal rights and options could make a voluntary
choice about whether to cooperate and answer police questions or to exercise
their constitutionally protected right to remain silent. Suspects unsure of what
their best option might be were given the right to consult with an attorney
before deciding whether to talk or not. With those rights in place, arrestees
would have a reasonable opportunity to make voluntary and intelligent deci-
sions about their best options in responding to police questioning. At least, that
was the world anticipated by the Supreme Court in its Miranda decision. The
resulting legal reality was, however, to fall far short of this.
Under the legal framework announced in Miranda and refined in subsequent
caselaw, an arrested suspect had only to invoke his right to remain silent or
his right to the assistance of counsel to bring interrogation to an immediate
halt (Michigan v. Mosley 1975, Edwards v. Arizona 1981). If the police fail to
honor the rights invocation and instead continue to question a suspect, any
resulting statements are not to be admissible by the prosecution in its case-
in-chief at trial. Under these rules, invocation of Miranda rights constitutes
a potent speech act through which the suspect in police custody becomes a
juridical actor – someone whose linguistic act of invocation constrains both
the interrogating police officers, who must immediately cease interrogation,
and later on the prosecuting attorney at trial, who is deprived of the evidence
of any incriminating statement made by the suspect after invocation. Powerful
word magic indeed.
Because of these significant legal consequences attendant upon invocation
of the Miranda rights, courts have had to determine exactly what kind of
speech acts will count as legally felicitous, thereby triggering these legal con-
straints on the police and prosecution. Unlike the minister at a wedding or the
juror announcing a verdict, however, suspects in police custody have no ready
script channeling what they must say in order exercise their rights. All that the
arrested person knows in this regard is the information given by the Miranda
warnings recited to him at the beginning of the questioning session. Those
warnings provide some information, albeit perhaps in stilted and unfamiliar
legal language, about the nature of one’s constitutional rights while in police
custody. The warnings, however, do not provide any guidance to the arrestee
on how to claim those rights. Under those circumstances, one might assume,
then, that courts would generously construe attempts by legally naïve arrestees
to invoke their rights to be efficacious (cf. Solan and Tiersma 2005: 53–72,
Ainsworth 1993: 292–315).
Instead, the Supreme Court held that attempts by arrestees to invoke Miranda
rights would be legally efficacious only if the invocations were made using
clear, unequivocal, and unambiguous language (Davis v. United States 1994).
If not, the attempted invocation of rights is treated as an infelicitous speech act
 linguistic ideology in American police interrogation law 5

wholly without legal consequences. Police need neither halt the interrogation
nor even attempt to clarify whether the suspect is or is not attempting to assert
his constitutional rights. Rather, the Court held that the police may continue
questioning as though nothing had happened, and any resulting statements
would be admissible at trial. In the Davis case, the arrested suspect being
questioned said to his interrogators at one point, ‘Maybe I should talk to a
lawyer’. The Supreme Court disallowed this as a legally effective invocation to
the right to counsel, noting that it appeared to be on its face to be nothing more
than an equivocal assertion of the right to speak with a lawyer, and approved
the actions of the interrogators in ignoring what Davis had said and continuing
the interrogation without determining whether he meant to assert his right to
consult with a lawyer. The Court noted that Davis had successfully claimed his
rights only when, later in the interrogation, he said to his interrogators, ‘I think
I want a lawyer before I say anything else’.
In the aftermath of the Davis case, reviewing courts have been required to
determine on a case by case basis whether an attempted invocation in a particu-
lar case is ambiguous or equivocal, and therefore not entitled to legal effect, or
sufficiently clear and unambiguous to count as a legally operative speech act. In
doing so, courts are making assumptions, albeit mostly unarticulated assump-
tions, about the nature of human communication. However, consideration of
the insights from linguistic discourse analysis – particularly sociolinguistics and
pragmatics – can expose the problematic nature of many judicial assessments
of whether particular invocations should be respected and accorded legal force.
Specifically, the insights of Gricean conversational implicature, coupled with
a sociolinguistically grounded recognition of the impact of power asymmetry
on speech register, illuminate the inadequacy of current legal doctrine and
practice both in light of what linguistics can tell us about how communication
works and in light of the legal system’s stated goal of admitting only voluntary
confessions into evidence.
H.P. Grice’s influential pragmatic framework on the significance of implica-
ture in communication serves to explain why ordinary sequential conversation,
which literally construed would seem to be a series of non-sequiturs, never-
theless is understood by its participants as a coherent, responsive discourse.
Grice posited that conversation is a rule-governed activity, with participants
interpreting responses as relevant to the context of the situation and to what
has been said earlier in the interchange (Grice 1989). For example, consider the
following short conversation between two university colleagues:
(1) Colleague 1: ‘Have you eaten?’
Colleague 2: ‘Oh, I have a one o’clock class…’
Colleague 1: ‘Another time, then.’
6 The international journal of speech, language and the law

Interpreted literally, this conversation is a disconnected series of statements

about the unrelated topics of food consumption, proximate class schedule,
and a non-specific reference to an indefinite point of time in the future. To
someone who lacked an understanding of the conventions of conversational
implicature – say, a Martian armed with a good English language grammar
and dictionary – this exchange would be incomprehensible nonsense. But,
in light of the Gricean maxim of relevance, these colleagues (and anyone else
overhearing this conversation) have no problem interpreting it as an invita-
tion to have lunch together, declined by the offeree in order to prepare for a
class, followed by a responsive suggestion by the offeror that the invitation
will be renewed at some unspecified future date. So natural are the operations
of these implicatures that both speakers are likely unaware of the fact that
they occurred at all. If asked later to report on the conversation, each is likely
to fill in the literal gaps in the conversation with the implied but unspoken
information conveyed – ‘Mary asked me to have lunch with her, but I told
her I couldn’t because I was prepping for class. We decided to do it some
other day’.
In order for Gricean implicature to result in successful communication, the
parties to the exchange must share frames of cultural meaning. In the given
example, for instance, the implicature depends on mutual understanding that
eating is a social activity that can be shared with others, that teachers ought
to prepare for classes that they are to teach, that professional obligations such
as teaching should trump the desire to engage in social interaction, that prof-
fered invitations refused for adequate but situation-specific reasons ought to
be renewed to show that no offense is taken at the refusal, etc. The point that
Gricean analysis drives home is that competent communicators construe their
conversational partners’ language as though it were intended to be relevant to
the social context in which it occurs and as responsive to what has been said
earlier in the conversation. In contrast, literal, acontextual interpretation of an
utterance would result in an inaccurate assessment of what a speaker means to
communicate in making that utterance.
Not only are utterances to be construed as relevant and responsive, but their
interpretation must also take into account the ways in which social identity and
social context shape speakers’ choices about how they will choose to express
themselves. Sociolinguistic research exploring the variability of linguistic
register used by speakers depending on those social factors helps to explain
systematic structural variation in word choice and syntax by speakers situated
in a particular cultural and social setting. Not every semantic imperative, for
example, is accomplished by speakers using bald and unmitigated syntactic
imperatives. Instead, speakers often use hedged or indirect language in prefer-
ence to an unmodified direct imperative. Furthermore, the likelihood that
 linguistic ideology in American police interrogation law 7

someone will express themselves using hedged or indirect language is increased

when there is power asymmetry between the parties, with the relatively power-
less speaker unlikely to make direct and unhedged demands upon the more
powerful party.
Robin Lakoff suggested that the use of hedged and indirect language to
express demands and requests might be seen as a characteristic of women’s lan-
guage (Lakoff 1973). Although later empirical research on gender and language
use has not fully confirmed Lakoff ’s assumptions about hedged language as a
gender-linked characteristic, this research has instead demonstrated a robust
link between relational powerlessness and the use of hedged and indirect
imperatives (Holmes 1984: 157, O’Barr and Atkins 1980). William O’Barr’s
study of language used by witnesses in court showed such a tight correlation
between the use of such language and being at a power disadvantage that he
came to call this speech register ‘powerless language’ (O’Barr 1982: 65–71,
Conley and O’Barr 2005: 63–64).
Being under arrest and questioned by the police is obviously a situation
in which there is an extreme power asymmetry between the interrogating
police officers and the suspect. Every aspect of the interaction underscores
that power imbalance. The physical setting and circumstances of custodial
interrogation magnify the helplessness of the arrested person. The arrestee is
alone and isolated, and may be confronted by several accusing officers at once.
He is physically restrained during the interrogation and can neither leave
the room nor control the timing and duration of the interaction. He cannot
choose when to eat or sleep or smoke or use the bathroom – all of those normal
human activities now are permitted at the sole discretion of his interrogators.
Control of the discursive aspects of the interrogation is likewise monopolized
by the police. The tempo and tone of the questioning is set by the questioners.
The police decide which topics will be considered relevant and pursued and
which will not. The assessment of whether responses are deemed satisfactory
or otherwise is a matter of exclusive police judgment, whereas the arrestee
cannot successfully impose his judgment as to the adequacy or appropriate-
ness of the questions posed by the police. In short, police interrogations are
extremely asymmetrical discursive interactions in which the interrogators wield
almost total power over every aspect of the questioning (Shuy 1998: 12–50). If
that extraordinary logistical power asymmetry were not enough, the suspect
must also be acutely conscious of the fact that, if his answers are deemed to be
displeasing or inadequate, serious consequences may follow in the future. He
may, rightly or wrongly, fear physical mistreatment at the hands of his jailors.
Even if he is not worried about that possibility, he must in any event certainly
consider that imprisonment or even death may result from the conviction that
his answers may facilitate.
8 The international journal of speech, language and the law

It was a recognition of the extreme power asymmetry inherent in custodial

police interrogation, coupled with its potentially devastating consequences
for arrestees who made incriminating statements under that interrogation,
that impelled the Supreme Court in Miranda to require the police to inform
arrested suspects that they had the constitutional right not to answer questions
and to seek the advice of counsel during interrogation as a way of ensuring
that the choice to answer police questions was in fact a free and voluntary one
(Miranda v. Arizona 1966). By invoking those rights, an arrestee could halt
an interrogation that threatened to overwhelm his ability to make free and
voluntary choices. However, the very fact that justifies Miranda warnings in the
first place – the coercive impact of police domination in the interrogation room
– also impacts the kind of language that an arrestee is apt to use in expressing
his intent to exercise those constitutional rights. The sociolinguistic research on
powerlessness and language strongly suggests that suspects undergoing police
interrogation would frequently attempt to invoke their rights using the hedged
and indirect language characteristic of ‘powerless language’. This is exactly what
an examination of the record in these cases confirms.
There have been several hundred state and federal cases adjudicating the
adequacy of an attempted invocation of Miranda rights in the wake of the
Supreme Court’s ruling in Davis v. United States. The vast majority of these
cases conclude that the invocation at issue was legally invalid. A look at this
post-Davis caselaw assessing arrestees’ attempts to exercise their Miranda
rights illustrates the variety of pitfalls that await suspects trying to claim their
constitutional rights. For example, some people made the mistake of trying to
invoke their right to a lawyer by asking for it using an interrogative syntactic
form instead of than an imperative one.
(2) a. ‘Could I call my lawyer?’ (Dormire v. Wilkinson 2001)
b. ‘Can I get my lawyer?’ (State v. Nixon 1996)
c. ‘Could I get a lawyer?’ (U.S. v. Wesela 2000)
d. ‘May I call a lawyer? Can I call a lawyer?’ (State v. Payne 2001)
e. ‘Do you mind if I have my lawyer with me?’ (U.S. v.
Whitefeather 2006)
f. ‘Can I speak to an attorney before I answer the question to find out
what he would have to tell me?’ (Taylor v. Carey 2007)
All of these requests were held to be disqualified as invocations because they
were interpreted to be merely questions about the theoretical availability of
counsel rather than as requests for counsel. In doing so, the courts parsed the
literal meaning of the arrestees’ utterances as though they were unrelated in
meaning to the preceding Miranda warnings telling them that they had the
right to assistance of a lawyer during questioning.
 linguistic ideology in American police interrogation law 9

Another mistake made by suspects was to use language that softens the
harshness of a bald demand for a lawyer. The following attempts to invoke the
right to counsel are examples of the kind of attempted invocations held to be
legal nullities because softening language was said to render them ambiguous
or equivocal:
(3) a. ‘I think I would like to talk to a lawyer.’ (Clark v. Murphy 2003)
b. ‘I think I will talk to a lawyer.’ (State v. Farrah 2006)
c. ‘It seems like what I need is a lawyer … I do want a lawyer.’
(Oliver v. Runnels 2006)
d. ‘Actually, you know what, I’m gonna call my lawyer. I don’t feel
comfortable.’ (People v. McMahon 2005)
Preceding the request with introductory qualifying language was fatal to
the suspect who said, ‘If I’m going to jail on anything, I want to have my
attorney present before I start speaking to you about whatever it is you guys
are talking about’ (Kibler v. Kirkland 2006). It didn’t matter to the court that
in fact the suspect was indeed going to jail as a result of the investigation in
progress; it was the very existence of any qualifying clause that doomed this
invocation.
In some cases, arrestees realized that they would be unable to contact an
attorney without logistical help from the police. Those who asked for police
assistance in reaching an attorney, however, found that their invocations were
invalidated as a result. For example, an arrestee who responded to the Miranda
warnings with a request that the police retrieve his lawyer’s business card had
not invoked his right to counsel, one appellate court decided (US v. Tran 2006).
Neither did a hospitalized suspect who asked the police, ‘Could I get a phone
in here so I can talk to a lawyer?’ (Jackson v. Commonwealth 2006).
Those who attempted to exercise their constitutional right to remain silent
fared no better than those trying to obtain a lawyer. The following responses to
the Miranda warnings were all been held too ambiguous or equivocal to count
as successful invocations of the right to silence:
(4) a. ‘I don’t want to talk about it.’ (Owen v. State 2003)
b. ‘I don’t have anything to say.’ (State v. Hickles 1996)
c. ‘I don’t got nothing to say.’ (U.S. v. Banks 1996)
d. ‘I don’t wanna talk no more.’ (U.S. v. Stephenson 2005)
e. ‘I just don’t think I should say anything.’ (Burket v. Angelone 2000)
f. Officer: ‘Do you want to make a statement to us?’ Arrestee: ‘Nope.’
(James v. Marshall 2003)
Nor is simply maintaining a stony silence in the face of police questioning
sufficient to count as a valid invocation of the right to remain silent, according
10 The international journal of speech, language and the law

to the Wisconsin Supreme Court (State v. Ross 1996). Ironically, the Court held
that, as a matter of law, a suspect has to explicitly speak in order to claim his
constitutional right to remain silent.
Even combining a refusal to answer questions with an explicit citation to
the constitutional provision incorporating the right to remain silent was not
enough for the prisoner who, at one point, said to the police, ‘I don’t want to
talk about this no more. That’s wrong, that’s wrong’. When this attempt to end
the questioning session was ignored by the police, who instead followed up
with a further barrage of questions and insinuations, the arrestee responded
with ‘I’m through with this, I’m through. I wanna be taken into custody, with
my parole…’ Again his attempt to cut off questioning was ignored by the police
who persisted in his questioning. The arrestee tried once again, responding to
the continued questioning by saying ‘I plead the Fifth’. The officer’s reply was,
‘Plead the Fifth. What’s that?’ and he carried on the interrogation unabated.
Amazingly, despite the fact that ‘taking the Fifth’ is a commonly used colloquial
expression in the United States to refer to exercising one’s right to remain silent
under the Fifth Amendment, this prisoner’s repeated attempts to cut off police
interrogation were held to be too ambiguous to be an effective invocation
(Anderson v. Terhune 2006).
Combining a refusal to answer questions with a request for a lawyer was also
not enough to trigger constitutional protections:
(5) a. ‘I don’t even want to talk unless I have me a lawyer and go through
this shit.’ (Harper v. State 2001)
b. ‘I don’t feel like I can talk with you without an attorney sitting right
here to give me some legal advice.’ (Baker v. State 2005)
c. ‘I’ll be honest with you. I’m scared to say anything without talking
to a lawyer.’ (Midkiff v. Commonwealth 1995)
d. Arrestee responded to police questions by saying, ‘You’re going to
have to ask my lawyer that.’ (U.S. v. Langford 2005)
e. Suspect responded to police questioning with, ‘Fuck you, talk to
my lawyer.’ (People v. Varnum 2004)
f. Arrestee responded to police officer saying, ‘Having these rights
in mind, do you wish to talk to us?’ with ‘Can I put ‘no’ ‘til I get
my lawyer?’ (State v. Jackson 2001)
These cases are among the most compelling for finding an invocation, in that
they exemplify the very concern that led the Supreme Court in Miranda to
interpose a right to counsel in the police interrogation context. As the Court
saw it, a legally naïve arrestee might well not be in a position to determine
how to respond to police questioning, or indeed whether to respond at
all, without the assistance of legal counsel to advise him about how best to
 linguistic ideology in American police interrogation law 11

protect his interests. Those suspects whose attempts at invocation expressly

articulate their need for legal advice before answering police questions thus
ought to be cases deserving the most generous construal of the adequacy of
rights invocations.
Gricean principles of conversational implicature suggest that courts should
interpret the kind of utterances seen in these cases as relevant and responsive
to the conversational situation at hand; namely, a police interrogation in which
the suspect has just been told that he has the right to refuse to answer ques-
tions and the right to the assistance of a lawyer during questioning. As Grice’s
framework and sociolinguistic research would predict, people frequently use
indirect and modified forms of imperatives that are nevertheless intended to
be interpreted as unequivocal demands. For example, just as a diner might
tell the waiter, ‘I think I would like to have the salmon’, or ‘Can I have another
cup of coffee?’ and would expect the waiter to understand those utterances as
unequivocal orders, so too an arrestee who says, ‘I think I would like to talk
to a lawyer’ (Clark v. Murphy 2003) or ‘Can I get my lawyer?’ (State v. Nixon
1996) should be understood as having made unambiguous requests for counsel.
A linguistically informed judge would recognize that speakers in powerless
positions are quite likely to resort to indirect and hedged syntactic forms
in lieu of using unmodified imperatives, but that utterances in a ‘powerless’
register are intended by their users to be no less unequivocal than those that
are syntactically more direct.
By situating the arrestees’ utterances into their discursive context, it becomes
clear that these responses by arrestees represent attempts, even if somewhat
indirect or inarticulate, to invoke rights guaranteed under the Constitution.
Yet, all too frequently, reviewing courts, applying hyper-literal readings to these
replies, have held that they did not constitute successful invocations of the right
to counsel or to remain silent. Instead of being legally operative requests, these
replies were deemed infelicitous legal speech acts because they took the form
of questions, or were framed in the subjunctive mood, or preceded the request
with softening expressions of emotion or desire.
The doctrinal message of the Supreme Court in Davis v. United States that
only clear, unambiguous, and unequivocal invocations are legally effective
has been treated by lower courts as a signal that they should indulge every
presumption against effective invocation as they interpret the language used by
arrestees attempting to claim their constitutional rights. As a result, unless the
right to counsel is invoked using a bald and completely unmodified impera-
tive such as ‘Give me a lawyer’, or ‘I want a lawyer’, the arrestee’s attempt to
exercise this constitutional right is unlikely to be respected either by the police
or by reviewing courts. Any arguable hedging or indirectness in locution by
a suspect attempting to invoke the Miranda rights will be construed as failing
12 The international journal of speech, language and the law

the requirement that invocation be clear, unambiguous, and unequivocal and

thus disqualified as a legally effective speech act. Ironically, these courts have
been even more hostile to finding legally efficacious invocation than was the
Supreme Court itself in the Davis case, which ultimately determined that Davis’
later statement, ‘I think I want a lawyer before I say anything else’, did constitute
an effective invocation of the right to counsel (Davis v. United States 1994). Few
reviewing courts today apparently agree.
Why, then, do appellate judges articulating the appropriate doctrinal rule and
trial judges applying the rule to the specific facts of particular police-suspect
interchanges disqualify so many of these attempts by people to invoke their
constitutional rights? One possibility is that these decisions reflect linguistic
ignorance on the part of judges and unawareness of how to infer utterance
meaning from context. However, a look at judges’ consideration of the kind
of language or behavior by suspects that will show that they have waived their
Miranda rights contrasts sharply with their assessment in cases where the ques-
tion is whether they have invoked them, suggesting that linguistic ignorance
does not account for these rulings.
The Miranda Court acknowledged that an arrestee might legitimately wish
not to exercise the right to remain silent and the right to counsel but rather
to waive those rights and voluntarily respond to police questioning, but it
noted that ‘a heavy burden rests on the government’ to show the validity of
any purported waiver (Miranda v. Arizona 1966: 475). The Court appeared to
anticipate that waiver be proven through specific acts or words of the suspect:
‘[A] valid waiver will not be presumed simply from the silence of the accused
after warnings are given or simply from the fact that a confession was in fact
eventually obtained’ (Miranda v. Arizona 1966: 475).
Later, however, the Supreme Court backed away from this position. In North
Carolina v. Butler, the arrestee was never asked whether he waived his rights
and never expressly did so. His only overt response to the Miranda warnings
was to refuse to sign a form indicating a waiver of his rights. Nevertheless,
the Supreme Court found that his silence in the face of the warnings, coupled
with the making of incriminating statements in response to police questioning,
constituted what the Court called ‘a course of conduct indicating waiver’ (North
Carolina v. Butler 1979: 373). Once the Butler opinion approved the finding
of this kind of implied waiver, it became unnecessary to point to any kind of
overt expression of understanding of rights or waiver by the arrestee. As long
as the officer testified that the rights were read to the suspect and the suspect
eventually made incriminating statements, that became enough to satisfy what
the Miranda Court had once called the ‘heavy burden’ on the prosecution to
show a knowing, voluntary, and intelligent waiver of rights.
 linguistic ideology in American police interrogation law 13

Examination of some representative post-Butler cases defining implied

waiver presents a stark contrast between the evidence sufficient to find a
rights waiver with the precision in language usage that courts require in
invocation cases. In one typical case, an arrestee refused to sign a waiver of
rights form when it was presented to him. The police then presented him
with false evidence purporting to implicate him in the crime. Not only did
the arrestee not have to use any particular, precise language in order to validly
give up his constitutional rights, but the mere fact that he made inculpatory
responses to the police lies was deemed to show an adequate waiver of his
Miranda rights (U.S. v. Middleton 2007). In a similar case, a defendant with
limited understanding of English was read a Vietnamese language version of
the Miranda warnings admittedly riddled with errors. After the police falsely
claimed they had evidence he had been seen at the scene of the crime, he
made incriminating statements. Despite the conceded defects in the warn-
ings he had been read and the fact that he never affirmatively waived his
rights in any way, he too was held to have validly waived his rights simply
by making incriminating statements (Thai v. Mapes 2005). The defendant in
another case merely nodded when the officer asked him if he understood
his rights. Although he was never asked to affirmatively waive his rights, his
responses to police interrogation were themselves held to be proof that he
waived his Miranda rights (Walkinhawk v. Pliler 2006). In another case, an
arrestee refused to sign a Miranda waiver form and twice explicitly stated
that he was not going to waive his rights. But, when he made inculpatory
responses to police questioning, the court had no trouble in concluding that
his explicit assertion that he did not intend to waive his rights was irrelevant.
The incriminating answers procured by the police themselves were proof
of a waiver of his rights (U.S. v. Acosta 2006). In yet another recent case, an
implied waiver of Miranda rights was found where the defendant’s ability to
write his name and answer questions was considered adequate evidence that
he had sufficient intelligence to understand that Miranda warnings, his book-
ing photographs failed to show evidence of gross physical mistreatment that
would raise questions as to the voluntariness of waiver, and his record of prior
convictions proof that he must have had ‘at least a rudimentary understanding
of his rights’ (U.S. v. Cuevas-Robledos 2006). This opinion flies in the face of
the Miranda Court’s express insistence that evidence of past encounters with
the police were inadequate to show appropriate knowledge of one’s rights,
since what if anything a suspect learned about the constitutional rights in
any earlier experience could ‘never be more than speculation’ (Miranda v.
Arizona 1966: 471–472).
In short, any time that the accused makes incriminating statements during
interrogation and fails to surmount the linguistic hurdles necessary to affirma-
14 The international journal of speech, language and the law

tively invoke his constitutional rights, most courts will find an implied waiver of
those rights. Far from being what the Miranda Court called a ‘heavy burden’ on
the state, waiver is now the default presumption which, as a practical matter, the
defendant must rebut by showing that he successfully navigated the linguistic
minefields of invocation law. Unless the accused affirmatively takes specific
steps to invoke his rights, any incriminating responses that he makes to police
questioning will be taken as conclusive evidence that he validly waived his
rights (Burket v. Angelone 2000: 198, U.S. v. Kaufman 2004: 256).
As a matter of public policy, it is surely debatable whether it is appropriate
to use such an expansive construal of the concept of waiver with respect to
legal rights as important as those enshrined in the constitution. What is clear,
however, is that it is impossible to conceptualize these cases as cases involving
waiver of rights if a literal reading of the suspect’s actual language were deter-
minative. Unless judges were at least intuitively aware of how to infer meaning
indirectly from context, they could not conclude that waivers existed in these
cases in which the suspects never explicitly acknowledge and give up their
rights. Whether or not these judges have a specific awareness of the principles of
Gricean implicature, in the waiver cases they appear quite capable of implicat-
ing non-literal meaning. So, general linguistic ignorance of implicature does
not appear to be an adequate explanation for the failure of these same judges
to give legal effect to most attempted Miranda rights invocations.
Instead, I will suggest that failure to give effect to most attempted invocations
results at least in part from an ideology of language that is deeply entrenched
within legal thought. Like most ideologies, it is largely invisible, seldom articu-
lated, and – if consciously thought of at all – considered to be purely common
sense. As a result, it is difficult to directly challenge language ideology. As John
Griffiths pointed out (1970: 359), ‘Ideological beliefs are pre-logical because
they determine the structure of perception and consciousness and therefore
are enmeshed in the factual and linguistic premises of argument’.
Judith Irvine (1989: 255) defines language ideology as ‘the cultural (or sub-
cultural) system of ideas about social and linguistic relationships, together with
their loading of moral and political interests’. Language ideology does not itself
produce patterns of language usage, but it does act to rationalize them and
render some patterns seemingly natural, and thus appropriate as the object of
preferential status. In the case of the ideology of language inherent to law, one
can see both factual claims and normative claims:
1 Language is a medium that maps human thought onto an objective
external reality and it does so transparently.
This belief in a correspondence theory of human communication acts as the
linchpin of legal ideology of language. This particular tenet of linguistic ideol-
 linguistic ideology in American police interrogation law 15

ogy is on occasion resisted by some judges, as when California Supreme Court

Justice Traynor derided his fellow judges’ ‘belief in the possibility of perfect
verbal expression…a belief (that) is a remnant of a primitive faith in the inher-
ent potency and inherent meaning of words’ (Pacific Gas and Electric Co. v.
G. W. Thomas Drayage and Rigging Co. 1968: 643–644). However, as Traynor
himself ruefully recognized, this belief is clearly the mainstream ideological
position in American legal doctrine and practice.
Probably the most celebrated articulation of this position in constitutional
law can be found in the writings of Supreme Court Justice Antonin Scalia, who
has forcefully argued that texts, in particular the Constitution, have a fixed,
determinate meaning. For Scalia, judges have an obligation to be faithful to
what he calls the ‘plain meaning of a text’, which he frequently identifies through
an appeal to linguistic authoritative sources such as dictionaries (Scalia 1989:
1184). Scalia’s version of textualism requires a commitment to a correspondence
theory of language.
Even if one were committed to a correspondence theory of language, one
might well grant that individuals have a privileged epistemological position
with respect to the existence or non-existence of their own mental states.
Nevertheless, because the mental states and processes of people are highly
pertinent to the resolution of many legal questions, legal ideology insists that
mental states are no different in their susceptibility to objective proof than
other aspects of reality:
2 A person’s own mental states – such as desire or intent – are a part of
externally verifiable objective reality.
As phrased by a judge in a nineteenth century case, ‘The state of a man’s mind
is as much a fact as the state of his digestion’ (Edgington v. Fitzmaurice 1882:
483). Consequently, the question of whether a person intended to exercise his
constitutional rights can be, and should be, resolved by an objective judicial
interpretation of the language used. That position, of course, entails the follow-
ing tenet in the linguistic ideology of the law:
3 Language is best understood objectively, from the point of view of
what a reasonable person would think something meant, not subjec-
tively from the point of view of idiosyncratic speakers and hearers.
Therefore, language is best understood by reference to authoritative
sources on objective meaning.
These factual commitments lead to the following normative commitments:
4 People have an obligation to use language transparently and bear the
responsibility if they fail to use language precisely and appropriately.
16 The international journal of speech, language and the law

5 Therefore, it is fair to hold people to the objective meaning entailed

by the language that they use, irrespective of what they subjectively
intended that language to mean.
These normative commitments have often been articulated in the realm of
contract law. For instance, the question of whether an enforceable contract
requires a ‘meeting of the minds’ between the parties to the contract or whether
courts instead should make an objective assessment of the language in question
has generally been answered by modern contract law in favor of the objective
theory of contract formation. To quote one typical case on this point:
[T]he inner intention of parties to a conversation subsequently alleged to
create a contract cannot either make a contract of what transpired or prevent
one from arising, if the words used were sufficient to constitute a contract.
In so far as their intention is an influential element, it is only such intention
as the words … of the parties indicate, not one secretly cherished which is
inconsistent with those words. (Embry v. Hargadine, McKittrick Dry Goods
Co. 1907: 778)
And, to continue in that vein, the esteemed Judge Learned Hand wrote,
If it were proved by twenty bishops that either party when he used the words
intended something else than the usual meaning which the law imposes on
them, he would still be held (bound to the contract). (Hotchkiss v. National
City Bank 1911: 293)
Based on that linguistic ideology, then, it is unsurprising that courts appear to
be less interested in how people actually express themselves in claiming their
rights than in imposing an explicitly normative regime on how people ought
to do so. Specifically, courts have been resistant to considering the implications
of research in pragmatics and sociolinguistics that empirically indicate how
language is used by people in trying to claim their constitutional rights in favor
of imposing an explicitly normative disciplinary rubric on how they ought to
use language in that context.
This linguistic ideology explains why courts seldom consider context in inter-
preting possible invocations and ignore the kind of conversational implicature
that Grice tells us in crucial in constructing meaning in interchanges. It also
explains why judges often appear unconcerned about taking into considera-
tion the social context in which speakers and hearers actually conduct their
communicative exchanges. Instead, the legal rules governing Miranda rights
are applied acontextually, without regard for either conversational implicature
or for the ways in which power asymmetry may affect register selection. Since
language is essentially a transparent and objective medium of communication,
 linguistic ideology in American police interrogation law 17

it is entirely fair to require speakers to conform to objective norms of com-

munication, as those norms are imagined by the courts.
This ideology also helps elucidate the otherwise puzzling treatment by courts
of instances in which invocations are found to be legally invalid despite the
fact that they were not intended to be equivocal by the speaker and were not
ambiguous from the point of view of the hearer. Even in these cases in which
the suspect intends to exercise his rights and the police officer understood him
as doing so, the attempted invocation may nevertheless be given no legal effect
upon judicial review as long as the invocation could in theory be interpreted to
be ambiguous. Such cases are unusual, since the interrogating officers almost
invariably testify that they did not understand the defendant’s language to be
an invocation.
Two cases, however, that do raise the issue were decided in 2002 by the state
Supreme Courts of Wisconsin and Virginia. In the Wisconsin case, the arrestee
responded to a detective’s request to make a written statement by saying twice,
‘I think maybe I need to talk to a lawyer’. Because an invocation of the right
to counsel requires an immediate halt to further interrogation, the detective
testified that he terminated the interview, ‘just to be on the safe side’, but a
second officer resumed the interrogation afterward (State v. Jennings 2002:
145). The Wisconsin Supreme Court overturned the trial court’s finding that
an unequivocal invocation had been made, holding that the proper analysis is
neither to look at subjective intent by the speaker nor to consider the actual
understanding by the hearer, but instead to make an objective interpretation of
the language in question. In the Virginia case, the arrestee interrupted a detec-
tive badgering him into admitting guilt and said, ‘Can I speak to a lawyer? I
can’t even talk to a lawyer before I make any comments or anything?’ (Redmond
v. Commonwealth 2002: 697). The Virginia Court of Appeals held this to be an
unequivocal invocation of the right to counsel on the part of the arrestee, and
noted that ‘without any doubt, the detective’s response to Redmond demon-
strates that he understood Redmond’s statements to be a request for counsel’
(Redmond v. Commonwealth 2001: 276). The Virginia Supreme Court reversed,
holding that the utterances in question as a matter of law were not an efficacious
invocation of the right to counsel, measured by an objective test, regardless
of the intent of the speaker or the understanding of the hearer (Redmond v.
Commonwealth 2002). As these cases illustrate, it doesn’t matter that the speaker
intended his utterance to be an invocation of his rights and it doesn’t matter
that the police officer hearing that utterance understood that the suspect was
in fact trying to invoke them. Because language is considered to be an objective
system of meaning, the courts are free to impose a meaning – the objectively
‘correct’ meaning – on the suspect’s utterance without regard for speaker intent
and hearer interpretation.
18 The international journal of speech, language and the law

The power of this linguistic ideology also helps to explain the disparate judi-
cial treatment of invocations of Miranda rights, which as shown is extremely
stingy, as contrasted with judicial considerations of waiver of Miranda rights,
which is breathtakingly expansive. Once one conceptualizes Miranda as
a framework for the positive creation of rights that otherwise would not
exist, then the linguistic ideology of law not surprisingly privileges direct
and unmodified speech registers that literally and precisely articulate the
parameters of those rights over those that do not. Accordingly, this unspoken
linguistic ideology makes it appear to judges that it is fair and reasonable that
people who fail to deploy language in a literal, transparent way should not
be permitted to claim the benefit of constitutional protections that must be
affirmatively called into being. On the other hand, lack of legal constraint on
the police, inaccurately termed ‘waiver’, is seen by judges as the default condi-
tion inhering in the police-citizen interaction unless and until the arrestee
changes the default situation by affirmatively invoking his rights. This would
be particularly true for judges whose political beliefs already predispose
them to a commitment to unfettered police power to investigate crime and a
hostility to constitutional constraints on that power. Not surprisingly, then,
such judges easily come to presume that no special language by arrestees
is needed to create ‘waiver’ of Miranda rights, since lack of constitutional
protection in police interrogation is seen as the default state of affairs, and a
desirable state of affairs to boot.
Law’s ideology of language thus stands the Miranda Court’s set of assump-
tions about police interrogation on its head. The Miranda Court imagined
that rights invocation would be generously construed by courts if the arrestee
‘indicated in any way’ that he desired to exercise his rights, and, as a corollary,
that the government would have a ‘heavy burden’ to demonstrate instead that
these rights had been waived (Miranda v. Arizona 1966: 444–445, 475). The
normative claims of the law’s ideology of language, however, impelled the
perhaps inevitable reversal of that set of presumptions. Since, according to
that ideology, people have an obligation to use language appropriately by being
objectively clear in the creation of legal rights, if they fail to do so, it is fair for
the law to sanction them for their failure. Hence it is unsurprising that ‘waiver’
is easily found, based not on the usual waiver formulation of ‘an intentional
relinquishment of a known right’ (Johnson v. Zerbst 1938: 464) but rather on
nothing more than the absence of ‘proper’ invocation.
Many legal commentators expected the Supreme Court to overrule Miranda
when they recently reconsidered the constitutional status of that case (Dickerson
v. United States 2000). To the surprise of many, the Court declined to do so
and reaffirmed its validity. What is left in place, however, is the worst of all
possible worlds: the Miranda opinion’s framework has been retained as a matter
 linguistic ideology in American police interrogation law 19

of doctrinal principle, but, as this look at the caselaw shows, it has become
an empty husk. Yes, citizens still have the constitutional rights outlined in
Miranda, but, as a practical matter, these rights are perilously easy to waive
and nearly impossible to actually invoke. The Supreme Court in Davis reas-
sured us that a person ‘need not speak with the discrimination of an Oxford
don’ in claiming one’s constitutional rights (Davis v. United States 1994: 459).
Subsequent appellate cases applying the Davis formula for assessing invocations
have made that assurance ring hollow. As this study shows, few Oxford dons
could successfully exercise their rights under the standards being applied in
American courts today.

About the author

Janet Ainsworth is Dean’s Distinguished Scholar and Professor of Law at Seattle
University School of Law.

Cases cited
Anderson v. Terhune, 467 F.3d 1208 (9th cir. 2006).
Baker v. State, 214 S.W. 3d 239 (Ark. S. Ct. 2005).
Burket v. Angelone, 208 F.3d 172 (4th cir. 2000).
Clark v. Murphy, 317 F.3d 1038 (9th cir. 2003).
Davis v. United States, 512 U.S. 452 (1994).
Dickerson v. United States, 530 U.S. 428 (2000).
Dormire v. Wilkinson, 249 F.3d 801 (2001).
Edgington v. Fitzmaurice, L.R. 29 Ch. Div. 459 (1882).
Edwards v. Arizona, 451 U.S. 477 (1981).
Embry v. Hargadine, McKittrick Dry Goods C., 105 S.W. 777 (1907).
Harper v. State, Tex. App. LEXIS 7497 (2001).
Hotchkiss v. National City Bank, 200 F. 287 (S.D. N.Y. 1911).
Jackson v. Commonwealth, 187 S.W.3d 300 (Ky. S. Ct. 2006).
James v. Marshall, 322 F.3d 103 (1st cir. 2003).
Johnson v. Zerbst, 304 U.S. 458 (1938).
Kibler v. Kirkland, U.S. Dist. LEXIS 55719 (D. N.Car. 2006).
Michigan v. Mosley, 423 U.S. 96 (1975).
Miranda v. Arizona, 384 U.S. 436 (1966).
20 The international journal of speech, language and the law

Midkiff v. Commonwealth, 462 S.E.2d 112 (Va. S. Ct. 1995).

North Carolina v. Butler, 441 U.S. 369 (1979).
Oliver v. Runnels, U.S. Dist. LEXIS 50704 (E. D. Ca. 2006).
Owen v. State, 862 So. 2d 687 (Fla. S. Ct. 2003).
Pacific Gas and Electric Co. v. G. W. Thomas Drayage and Rigging Co., 442 P.2d 641 (Cal. S.
Ct. 1968).
People v. McMahon, 31 Cal. Rptr.3d 256 (Cal. App. 2005).
People v. Varnum, 2004 Cal. App. LEXIS 5189.
Redmond v. Commonwealth, 2001 Va. App. LEXIS 276.
Redmond v. Commonwealth, 568 S.E.2d 695 (Va. S. Ct. 2002).
State v. Farrah, Minn. App. Unpub. LEXIS 984 2006).
State v. Hickles, 929 P.2d 141 (Kan. S. Ct. 1996).
State v. Jackson 19 P.3d 121 (Kan. S. Ct. 2001).
State v. Jennings, 647 N.W.2d 142 (Wisc. S.Ct. 2002).
State v. Nixon, 687 So.2d 114 (La. Ct. App. 1996).
State v. Payne, 833 So.2d 927 (La. S. Ct. 2001).
State v. Ross, 552 N.W.2d 428 (Wisc. S. Ct. 1996)
Taylor v. Carey, U.S. Dist. LEXIS 12686 (E.D. Ca. 2007).
Thai v. Mapes, 412 F.3d 970 (8th cir. 2005).
U.S. v. Acosta, 363 F.3d 1141 (11th cir. 2006).
U.S. v. Banks, 78 F. 3d 1190 (7th cir. 1996).
U.S. v. Cuevas-Robledos, 2006 U.S. Dist. LEXIS 76300 (D. Oregon 2006).
U.S. v. Langford, 155 Fed. App’x. 936, U.S. App. LEXIS 24861 (8th cir. 2005).
U.S. v. Middleton, 2007 U.S. App. LEXIS 14293 (11th cir. 2007).
U.S. v. Stephenson, 152 Fed. App’x. 904 (11th cir. 2005).
U.S. v. Tran, 171 Fed. App’x. 758, U.S. App. LEXIS 5068 (11th cir. 2006.)
U.S. v. Wesela, 223 F.3d 656 (7th cir. 2000).
U.S. v. Whitefeather, U.S. Dist. LEXIS 17239 (D. Minn. 2006).
Walkinhawk v. Pliler, 184 Fed. App’x 653 (9th cir. 2006).
 linguistic ideology in American police interrogation law 21

References
Ainsworth, J. (1993) In a different register: the pragmatics of powerlessness in police
interrogation. Yale Law Journal 103: 259–322.
Austin, J. L. (1962) How to do Things with Words. (Second edition.) Cambridge, MA:
Harvard University Press.
Conley, J. M. and O’Barr, W. M. (2005) Just Words. (Second edition.) Chicago, IL:
University of Chicago Press.
Grice, P. (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press.
Griffiths, J. (1970) Ideology in criminal procedure. Yale Law Journal 79: 359–417.
Holmes, J. (1984) ‘Women’s language’: A functional approach. General Linguistics 24:
149-178.
Irvine, J. (1989) When talk isn’t cheap: language and political economy. American
Ethnologist 16: 248–267.
Lakoff, R. (1973) Language and women’s place. Language in Society 2: 45–80.
O’Barr, W. and Atkins, B. (1980) Women’s language or powerless language. In S.
McConnell-Ginet, R. Borker and N. Furman (eds) Women and Language in Literature
and Society 98–110. New York: Praeger.
O’Barr, W. (1982) Linguistic Evidence: language, power, and strategy in the courtroom. New
York: Academic Press.
Ogden, C. K. and Richards, I. A. (1989) The Meaning of Meaning. Orlando, FL: Harcourt
Brace Jovanovich.
Scalia, A. (1989a) The rule of law as a law of rules. University of Chicago Law Review 56:
1175–1188.
Shuy, R.W. (1998) The Language of Confession, Interrogation, and Deception. Thousand
Oaks, CA: Sage Publications.
Solan, L. M. and Tiersma, P. M. (2005) Speaking of Crime: the language of criminal justice.
Chicago: University of Chicago Press.