Elliott v. State

FINAL COPY




                    S18A1204. ELLIOTT v. THE STATE.


      PETERSON, Justice.

      The State is prosecuting Andrea Elliott for driving under the influence of

alcohol. When Elliott was arrested, she refused to submit to a breath test.

Georgia statutes allow the State to use her refusal against her in her criminal

trial, and the State has sought to do precisely that. The United States Supreme

Court has held that the Fifth Amendment to the United States Constitution does

not bar the State from using such a refusal, in part because the Fifth Amendment

gives Elliott no right to refuse to act in the first place. But we have held — and

hold again today — that the protection against compelled self-incrimination

provided by Article I, Section I, Paragraph XVI of the Georgia Constitution

does afford the right to refuse such a test. So Elliott argues to us that Paragraph

XVI gives her the protection that the Fifth Amendment does not, and thus

renders invalid the portions of the statutes allowing her refusal to be admitted

against her. We agree.
       In Olevik v. State, 

 (

) (2017), we held that —

unlike the Fifth Amendment — the Georgia Constitution’s right against

compelled self-incrimination prevents the State from forcing someone to submit

to a chemical breath test. Given that holding, and relying on a combination of

decisions by this Court and Fifth Amendment decisions of the United States

Supreme Court, Elliott argues that assertions of that right cannot be admitted

against her. The State argues that we were wrong in Olevik and should overrule

it, but also argues that the right — properly understood — does not prohibit the

State’s use of refusal evidence against a defendant.1 We adhere to Olevik, and,

after extensive review of the historical record and our case law, conclude that

our state constitutional right does prohibit admission of evidence that Elliott

refused a breath test.

       After a review of the undisputed facts, we begin by reviewing three

principles that guide our constitutional interpretation in this case. Turning then

to the State’s argument that we should overrule Olevik, a careful application of

those interpretive principles leads us to adhere to Olevik. We properly


       1
        The efforts of both parties to this case were augmented by various amici curiae. We
appreciate the helpful briefing.

                                            2
interpreted Paragraph XVI in the light of this Court’s consistent holdings that

Paragraph XVI’s materially identical precursors protected defendants from

being compelled to perform affirmative acts. We then review the relevant

history and case law regarding admission of a defendant’s refusal to act and the

drawing of adverse inferences therefrom. We conclude that although the pre-

Revolution English common-law right against self-incrimination did not

preclude admission of defendants’ refusals to incriminate themselves or adverse

inferences therefrom, legal developments in the United States and specifically

in Georgia in the years leading up to and around the time of the adoption of the

1877 Constitution demonstrate that the original public meaning of the 1877

Constitution’s precursor to Paragraph XVI (the “1877 Provision”) did preclude

the admission of such evidence. Finding no basis to conclude that the meaning

of Paragraph XVI itself as adopted in 1983 is different in that respect, we

conclude that the Georgia statute permitting admission of Elliott’s refusal

violates Paragraph XVI.

      I.    Background.

      The relevant facts are not in dispute. In August 2015, a police officer

stopped Elliott after observing her commit several traffic violations, including

                                       3
a failure to maintain her lane. During the stop, Elliott admitted to consuming

alcohol earlier that day. After smelling the odor of alcohol and observing several

signs of impairment, including several clues during a field sobriety test, the

officer arrested Elliott for DUI and other traffic offenses and read her the

statutorily mandated implied consent notice. See OCGA § 40-5-67.1 (b).2 Elliott

replied that she was overwhelmed and unsure of what was happening, so the

officer explained why he stopped her, why he asked her to perform field sobriety

tests, why he read her the implied consent notice following her arrest, and that

a refusal to submit to a state-administered breath test could result in certain

consequences, including that her refusal to submit might be offered into

       2
          The language of the implied consent notice for drivers aged 21 years or older (like
Elliott) is as follows:
        Georgia law requires you to submit to state administered chemical tests of your
        blood, breath, urine, or other bodily substances for the purpose of determining
        if you are under the influence of alcohol or drugs. If you refuse this testing,
        your Georgia driver’s license or privilege to drive on the highways of this state
        will be suspended for a minimum period of one year. Your refusal to submit
        to the required testing may be offered into evidence against you at trial. If you
        submit to testing and the results indicate an alcohol concentration of 0.08
        grams or more, your Georgia driver’s license or privilege to drive on the
        highways of this state may be suspended for a minimum period of one year.
        After first submitting to the required state tests, you are entitled to additional
        chemical tests of your blood, breath, urine, or other bodily substances at your
        own expense and from qualified personnel of your own choosing. Will you
        submit to the state administered chemical tests of your (designate which tests)
        under the implied consent law?
OCGA § 40-5-67.1 (b) (2).

                                             4
evidence against her at trial. Elliott refused to submit to a breath test and was

taken to jail. She filed a motion to suppress her refusal to submit to a breath test,

claiming that the introduction of that evidence at trial would violate her right

against compelled self-incrimination under the Georgia Constitution and

Georgia Code. The trial court denied her motion, leading to this appeal.

      II.    Principles of Georgia constitutional interpretation.

      Both Elliott’s arguments challenging the denial of her motion to suppress

evidence of her refusal and the State’s arguments that we should reconsider our

decision in Olevik require us to begin by reviewing some important principles

that guide our interpretation of the Georgia Constitution in this case.

      We have often explained that we interpret the Georgia Constitution

according to its original public meaning. And, of course, the Georgia

Constitution that we interpret today is the Constitution of 1983; the original

public meaning of that Constitution is the public meaning it had at the time of

its ratification in 1982. But many of the provisions of the Constitution of 1983

first originated in an earlier Georgia Constitution; unlike the United States, the

State of Georgia has had ten constitutions since declaring independence from



                                         5
Great Britain.3 The meanings of those previous provisions are critical to

understanding the meaning they carried at the time they were readopted. See

Clarke v. Johnson, 

 (

) (1945) (“A constitutional

provision must be presumed to have been framed and adopted in the light and

understanding of prior and existing laws and with reference to them.

Constitutions, like statutes, are properly to be expounded in the light of

conditions existing at the time of their adoption.” (citation and punctuation

omitted)). Paragraph XVI first appeared in the Constitution of 1877, and was

carried forward without material change into the Constitutions of 1945, 1976,

and now our current Constitution of 1983. Our focus on the original public

meaning of this provision thus requires us to consider two interpretive principles

that arise from the provision’s multi-constitutional history, and a third principle

that simply arises from the independent nature of state constitutions.



      3
         This number does not include the 1776 Rules and Regulations of the Colony of
Georgia, the first governing document during the revolutionary period (although adopted
three months before the signing of the Declaration of Independence). The first proper
constitution was adopted a year later in 1777. See Walter McElreath, A Treatise on the
Constitution of Georgia 67 (Atlanta: Harrison Co., 1912); see also Reheis v. Baxley
Creosoting & Osmose Wood Preserving Co., 

 (2) n.12 (

)
(2004). The Constitution of 1777 was followed by the Constitutions of 1789, 1798, 1861,
1865, 1868, 1877, 1945, 1976, and 1983.

                                          6
       A.     The presumption of constitutional continuity.

       Original public meaning is an interpretive principle that we apply to each

of our constitutions. See Padelford, Fay & Co. v. Mayor and Aldermen of the

City of Savannah, 

 (1854) (“[T]he Constitution, like every other

instrument made by men, is to be construed in the sense in which it was

understood by the makers of it at the time when they made it. To deny this is to

insist that a fraud shall be perpetrated upon those makers or upon some of

them.” (emphasis in original));4 see also Olevik, 

 (2) (c) (i)


       4
          Since the people are the ultimate “makers” of the Georgia Constitution, this requires
a focus on the public meaning, not the subjective intent of the drafters. See Olevik, 

 (2) (c) (i). All of our constitutions were the result of voter ratification or a
constitutional convention whereby delegates were elected by the voters to represent them.
The Constitutions of 1983, 1976, 1945, 1877, 1868, 1865, and 1861 were adopted upon voter
ratification after the General Assembly or a convention proposed a new constitution. See Ga.
L. 1983, p. 5188 (noting that constitution proposed by General Assembly, Ga. L. 1981, Ex.
Sess., p. 143, was approved by a vote of 567,663 to 211,342); Ga. L. 1977, p. 1897 (1976
Constitution proposed by the General Assembly, see Ga. L. 1976, p. 1198, § 1, was approved
by vote of 610,516 to 394,734); Ga. L. 1945, p. 8 (General Assembly proposed constitution
expressly requiring voter ratification); Ga. Const. 1877, Art. XIII, Sec. II, Pars. I & II (1877
Constitution to be adopted upon voter ratification); McElreath, supra, § 115 (“The
Constitution [of 1868] was adopted by a vote of 89,007 to 71,309); id., § 107 (1865
Constitution “submitted to the people for ratification and was adopted”); Ga. Const. of 1861,
Art. V, Sec. VII (constitution “shall not take effect until the same is ratified by the people”).
For the Constitutions of 1798 and 1777, voters elected delegates to attend a constitutional
convention that created and adopted those constitutions. See McElreath, supra, §§ 53, 57, 77.
The 1789 Constitution was the result of three conventions, at least one of which was
composed of three delegates from each county chosen by voters. Id., supra, § 72. Of course,
we do not consider the subjective intent of individual lawmakers even when the electorate’s
role is more attenuated. See Olevik, 

 (2) (c); Merritt v. State, 

                                 7
(citing cases). Because the meaning of a previous provision that has been

readopted in a new constitution is generally the most important legal context for

the meaning of that new provision, and because we accord each of those

previous provisions their own original public meanings, we generally presume

that a constitutional provision retained from a previous constitution without

material change has retained the original public meaning that provision had at

the time it first entered a Georgia Constitution, absent some indication to the

contrary. See, e.g., Lathrop v. Deal, 

 (III) (B) (

) (2017) (interpreting Art. I, Sec. II, Par. V of the Constitution of 1983 in the

light of the original meaning of the provision as it first appeared in the

Constitution of 1861); Ga. Motor Trucking Assn. v. Ga. Dept. of Revenue, 

 (2) (B) (

) (2017) (interpreting provision of the

Constitution of 1983 in the light of the meaning of amendments to the

Constitution of 1945 that were carried forward); Smith v. Baptiste, 

 (1) (

) (2010) (considering meaning of 1983 provision in part

in the light of the meaning of its predecessor provisions); 

 (2)-(3)



656-657 (

) (2010) (Nahmias, J., concurring specially) (explaining difficulty in
attempting to determine intent of legislatures).

                                            8
(Nahmias, J., concurring) (same); Nelms v. Georgian Manor Condo. Assn., 

 (2) (

) (1984) (considering history of 1877 provision

carried forward in the 1945 and 1976 Constitutions, and concluding that the

provision’s meaning remained unchanged “[a]s the language of this paragraph

remained unchanged in the 1976 Constitution”); Bloomfield v. Liggett & Myers,

Inc., 

 (

) (1973) (interpreting Art. I, Sec. I,

Par. IV of the Constitution of 1945 in the light of the original meaning of the

same provision in the Constitution of 1877);5 cf. Bibb County v. Hancock, 

 (1) (

) (1955) (“It is clear that, in placing [a provision

from the Constitution of 1877] in the Constitution of 1945, there was no

intention to declare any new principle of law, but merely to continue in the new

Constitution the same provision of the old, with the same meaning.”); cf. State

v. Central of Ga. R. Co., 

 (

) (1900) (absent

constitutional text making clear a provision was intended to change the law,

       5
        Bloomfield has since been criticized as misstating the original meaning of the 1877
provision. See Baptiste, 

 (Hunstein, C. J., dissenting) (arguing that
Bloomfield made historical errors); but see 

 (2)-(3) (Nahmias, J., concurring)
(arguing Bloomfield correctly stated the original meaning of the 1877 provision). But
whether Bloomfield was right or wrong about the meaning of the 1877 provision is beside
the point we make here. Here, the only point for which we consider Bloomfield is its
unchallenged holding that the meaning of the 1877 provision — whatever that might be —
was the meaning of the same provision carried forward into the 1945 Constitution.

                                            9
constitutional text should be interpreted consistent with the common law that

preceded it). This presumption of constitutional continuity helps maintain the

stability of Georgia’s constitutional law, while still yielding when other

considerations make clear that the people have changed the meaning of a

provision.

      B. A constitutional clause that is readopted into a new constitution and
that has received a consistent and definitive construction is presumed to carry
the same meaning as that consistent construction.

      A second interpretive implication arises from applying an original public

meaning analysis to a constitutional provision that has been readopted without

material change in multiple constitutions, and this principle involves our

interpretations of the previous constitutions. In Olevik, we reasoned that

“history compel[led] our conclusion” that Paragraph XVI protected individuals

from being forced to perform incriminating acts. 

 (2) (c). That

history revealed a consistent and definitive construction of a self-incrimination

clause whose words remained materially unchanged since the clause first

appeared in our constitution in 1877. See 

 (2) (c) (ii). Given this

consistent and definitive construction, we presumed that construction was

carried forward into the 1983 Constitution. 

 (2) (c) (ii).

                                        10
      We have long applied this interpretative approach. In McKnight v. City

of Decatur, 

 (2) (

) (1946), we noted that “[t]he

framers of the revised Constitution [of 1945] were presumably cognizant of the

foregoing provisions of the earlier Constitutions [of 1868 and 1877, as amended

in 1919], and of the interpretations which this court had placed upon them.” Of

course, the relevant question is not the fact of what specific legislators did or did

not know; rather, the point is what was sufficiently part of the public legal

context such that a presumption is appropriate.

      Soon after our decision in McKnight, we explained the presumption that

the framers of a new constitution are not only aware of the provisions of the

earlier constitution, but when

      [they] adopt provisions contained in a former Constitution, to which
      a certain construction has been given, [they] are presumed as a
      general rule to have intended that these provisions should have the
      meaning attributed to them under the earlier instrument.

Thompson v. Talmadge, 

 (2) (

) (1947). And we

have applied this principle in many cases since. See, e.g., Atlanta Indep. School

Sys. v. Lane, 

 (2) (

) (1996); City of Thomaston v.

Bridges, 

 (

) (1994); Toombs County v. O’Neal, 254


                                         

, 391-392 (2) (

) (1985); Aldrich v. State, 

(

) (1964); Hancock, 

 (1); Griffin v. Vandegriff, 

 (

) (1949); see also McCafferty v. Med. College of

Ga., 

 (

) (1982) (Gregory, J., concurring specially),

overruled on other grounds by Self v. City of Atlanta, 

 (1) (

) (1989) (adopting special concurrence).

      Many of our sister states apply a similar principle. See, e.g., Fla. Dept. of

Revenue v. City of Gainesville, 918 S2d 250, 263-264 (Fla. 2005) (adopting

prior construction given to prior constitutional provision where framers of new

constitution could have changed the meaning by redefining terms or using

different terms altogether); Succession of Lauga, 624 S2d 1156, 1165 (II) (A)

(La. 1993) (“When a constitutional provision is identical or very similar to that

of a former constitution, it is presumed that the same interpretation will be given

to it as was attributed to the former provision.”); State ex rel. Ashcroft v. Blunt,

 (Mo. 1991) (same); Paper Supply Co. v. City of Chicago,

 (Ill. 1974) (prior construction applies unless “it is apparent that

some other meaning was intended”); Richardson v. Hare, 

(Mich. 1968) (same); First Trust Co. of Lincoln v. Smith, 

, 773

                                        12
(Neb. 1938) (“[W]hen the form of words used in the Constitution is borrowed

from an older source, it comes laden with its previous meaning.” (citation and

punctuation omitted)); Williamson v. City of High Point, 

(N.C. 1938) (where provision reincorporated into new or revised constitution,

“it will be presumed to have been retained with a knowledge of the previous

construction, and courts will feel bound to adhere to it” (citation and

punctuation omitted)); Ex parte Western Union Telegraph Co., 76 S 438, 439

(Ala. 1917) (“bound” by former construction given to a constitutional provision

that was readopted into new constitution); Elliott v. Ashby, 

 (Va.

1905) (“We have, then, a judgment of this court . . . construing a provision of

the Constitution and a statute passed in pursuance thereof, and the construction

placed upon them has not only been unchallenged, but must be taken as having

been approved by the adoption of the identical language in the Code and the

Constitution.”); Cline v. State, 

, 728 (Tex. Crim. App. 1896)

(constitutional provision with a settled judicial construction is presumed to have

that meaning when it is incorporated into new constitution).

      The State takes issue with this principle, arguing that its implications

would prevent us from ever reconsidering previous decisions no matter how

                                       13
wrong. But our longstanding articulation of this principle is not so restrictive.

Indeed, in one of the first cases referring to this principle, we expressly rejected

the notion that we could never reconsider such precedent. See Thompson, 

 (2) (observing that a rule that precludes a court from revisiting a

previous construction is “stated more strongly than we would be willing to put

it”). Instead, we explained that the better rule was that “where a constitutional

provision has received a settled judicial construction, and is afterwards

incorporated into a new or revised constitution, or amendment, it will be

presumed to have been retained with a knowledge of the previous

construction[.]” 

 (citation and punctuation omitted; emphasis supplied). The

same rule applies to statutes. 

 (2) (“The language of an existing statute

adopted into a constitution is presumed to be taken with its established

construction.”); see also Bradshaw v. State, 

 (2) (

) (2015) (noting that “when judicial interpretations have settled the meaning

of an existing statutory provision, repetition of the same language in a new

statute indicates, as a general matter, the intent to incorporate its judicial

interpretations as well” (quoting Jerman v. Carlisle, McNellie, Rini, Kramer &

Ulrich, L.P.A., 

 (

, 176 LE2d 519) (2010)

                                        14
(punctuation omitted))). The presumption created by a consistent and definitive

construction “reflects the value of consistency in the interpretation of legal

language.” Bridges, 

.6

       The State also argues that this principle should apply only when

constitutional text is ambiguous, citing statements from this Court more than

166 years ago that the first rule of construction is that an unambiguous statute

or provision “stands self-interpreted, and Courts have nothing to do but to

enforce it.” Neal v. Moultrie, 

 (1852). The State also notes

similar language in Griffin. See 

 (1).

       The State is wrong; when we determine the meaning of a particular word

or phrase in a constitutional provision or statute, we consider text in context, not

in isolation. See Chan v. Ellis, 

 (1) (

) (2015)

(“The common and customary usages of the words are important, but so is their

context.” (citation omitted)); Brown v. State, 

 (2) (b) (725



       6
        The presumption arising from a consistent and definitive construction, however, like
most legal presumptions, may be rebutted. As we explain later, this presumption is not
rebutted as we applied it in Olevik, and no such construction exists as to the refusal issue.
Accordingly, this is not a case that calls us to articulate precisely when such a presumption
may be rebutted.


                                             

) (2012) (“words often gain meaning from context”); see also Upper

Chattahoochee Riverkeeper, Inc. v. Forsyth Cty., 

 (1)

(

) (2012) (“[E]ven if words are apparently plain in meaning, they

must not be read in isolation and instead, must be read in the context of the

regulation as a whole.”). For context, we may look to “the broader context in

which that text was enacted,” including other law — constitutional, statutory,

decisional, and common law alike — that forms the legal background of the

constitutional provision. Olevik, 

 (2) (c) (i); see also Undisclosed

LLC v. State, 

 (2) (a) (

) (2017); Gilbert v.

Thomas, 

 (1847) (rejecting interpretation of constitutional text

read in isolation because broader legal context showed that it was not “the

design of the framers of the constitution”). The presumption arising from a

consistent and definitive construction is simply a reflection of this principle.

      C. Georgia constitutional provisions may confer greater, fewer, or the
same rights as similar provisions of the United States Constitution, and
decisions of the United States Supreme Court interpreting those similar
provisions are persuasive in our interpretation of the Georgia Constitution only
to the extent that those decisions are rooted in shared history, language, and
context.




                                       16
      When we consider the meaning of a provision of the United States

Constitution, we faithfully apply the decisions of the United States Supreme

Court as to the meaning of that provision. Such a faithful application is not an

act of judgment on our part; it is an act of obedience. But when the provision we

consider is a provision of the Georgia Constitution, our approach is different.

      When interpreting a provision of our Constitution that parallels a

provision of the United States Constitution, we should take seriously decisions

of the United States Supreme Court that have interpreted that parallel provision.

And here, the federal self-incrimination clause of the Fifth Amendment, see U.

S. Const. amend. V (“No person . . . shall . . . be compelled in any criminal case

to be a witness against himself[.]”), is similar to the state self-incrimination

clause of Paragraph XVI (“No person shall be compelled to give testimony

tending in any manner to be self-incriminating.”). But we owe those federal

decisions no obedience when interpreting our own Constitution. “Questions of

the construction of the State Constitution are strictly matters for the highest

court of this [S]tate. The construction of similar federal constitutional

provisions, though persuasive authority, is not binding on this [s]tate’s



                                       17
construction of its own Constitution.” Pope v. City of Atlanta, 

(1) (

) (1977). And so it follows, as we observed in Olevik:

      State constitutional provisions may, of course, confer greater

      protections than their federal counterparts, provided that such

      broader scope is rooted in the language, history, and context of the

      state provision. In the same way, a state constitution may also offer

      less rights than federal law, so long as it does not affirmatively

      violate federal law.

Olevik, 

 n.3 (citations omitted).

      This is scarcely a Georgia-specific idea. State constitutional rights were

“meant to be and remain genuine guarantees against misuse of the state’s

governmental powers, truly independent of the rising and falling tides of federal

case law both in method and in specifics.” State v. Kennedy, 

 (Or. 1983). See also Massachusetts v. Upton, 

 (

, 80 LE2d 721) (1984) (Stevens, J., concurring in the judgment); State v.

Walker, 

 (Utah 2011) (Lee, J., concurring); Malyon v.

Pierce County, 

, 1281 n.30 (Wash. 1997); Sitz v. Dept. of State

Police, 

 (Mich. 1993); Ex parte Tucci, 

, 32

                                       18
n.34 (Tex. 1993) (Phillips, C. J., concurring); State v. Smith, 

(Or. 1986) (plurality); Jeffrey S. Sutton, 51 Imperfect Solutions: States and the

Making of American Constitutional Law 174-178 (2018).

      And so, as we have previously explained, any decision about the scope of

a provision of the Georgia Constitution must be “rooted in the language, history,

and context” of that provision. Olevik, 

 (2) (b) n.3 (citing Grady

v. United Govt. of Athens-Clarke County, 

 (2) (b) (

) (2011)). Many of our constitutional provisions will find their roots in

language, history, and context shared by similar provisions of the United States

Constitution. And when the United States Supreme Court interprets such a

provision, its decision may well shed useful light on the language, history, and

context that should guide our interpretation of the similar Georgia provision.

But decisions of the United States Supreme Court interpreting similar provisions

generally will prove persuasive only to the extent that the Court’s decisions

actually were guided by that same language, history, and context.

      With these guiding principles in mind, we should not simply recite

holdings of the United States Supreme Court regarding the Fifth Amendment

and uncritically import them into our interpretation of Paragraph XVI of the

                                       19
Georgia Constitution. Rather, we must first discern the original public meaning

of Paragraph XVI in order to address the merits of the parties’ arguments. See

Olevik, 

 (2) (c) (i) (“We interpret a constitutional provision

according to the original public meaning of its text, which is simply shorthand

for the meaning the people understood a provision to have at the time they

enacted it.”). But Paragraph XVI is a holdover from prior constitutions, and we

must remember that the right against compelled self-incrimination was first

constitutionalized with the adoption of the 1877 Constitution, thereby

preserving a right at common law. In order to understand the meaning of the

constitutional right, we must understand the nature of that common-law right as

it was understood in 1877. We must also determine if a consistent and definitive

construction of the constitutional right now found in Paragraph XVI applies to

the issues raised here, and, if so, whether we should presume that construction

was carried forward into the Constitution of 1983.

      III.   We adhere to our decision in Olevik.

      Olevik posed the question of whether Paragraph XVI’s mandate that “(n)o

person shall be compelled to give testimony tending in any manner to be self-

incriminating” applied to chemical breath tests. 302 Ga. at 235 (2) (c). In

                                      20
answering that question, we reviewed an unbroken line of precedent beginning

with Day v. State, 

 (1879), and clarified in Calhoun v. State, 

 (

) (1916), that interpreted the 1877 Provision that is virtually

identical to Paragraph XVI of the 1983 Constitution as barring compelled acts,

not merely compelled oral or written testimony. Based on the well-established

meaning given to the constitutional right against compelled self-incrimination

and carried forward into subsequent state constitutions, we concluded that a

breath test is an act incriminating in nature and, therefore, Paragraph XVI

prohibits the State from compelling such a test. Olevik, 

 (2)

(c).

       Unhappy with our decision in Olevik and its potential implications, the

State has, in this case and in other appeals currently before the Court, asked us

to reconsider our decision in Olevik. In this case, the State does not challenge

our conclusion in Olevik that the sustained strong blowing necessary for a

breath test properly is analyzed as an affirmative act. Instead, the State

challenges the holding itself, arguing that this Court’s decisions in Day and

Calhoun were wrong because they contravened the plain meaning of the word

“testimony” found in the constitutional self-incrimination provision, and

                                       21
because their reasons for expanding the scope of the right to include affirmative

acts were based on a misunderstanding of the common-law right from which the

constitutional right is derived. The State also argues that this Court stands alone

in affording protection to incriminating acts, and we should start over with our

interpretation of Paragraph XVI according to the plain meaning of its text —

without considering the prior constructions given to the right against compelled

self-incrimination — because the 1983 Constitution was a “new” constitution

that was meant as a departure from established jurisprudence. The State

concludes that a plain reading of Paragraph XVI would limit the right against

compelled self-incrimination to only oral and written testimony, which would

be coterminous with the Fifth Amendment’s self-incrimination right.

       Undertaking that review, we conclude that, although the scope of the

common-law right against compelled self-incrimination is indeed debatable, our

case law around the time the 1877 Provision was adopted had extended the right

to more than oral and written testimony. That case law included our decision in

Day, which research reveals was not out of step with other courts that

considered the issue around the same time. Indeed, the right against compelled

self-incrimination was previously understood in many courts, including federal

                                        22
courts, to be more expansive than it is now. The fact that most of those

jurisdictions have since narrowed the scope of the right against compelled self-

incrimination does not require that we do the same for Paragraph XVI. We have

consistently and definitively construed the right against compelled self-

incrimination to bar compelled acts, and there is nothing about Paragraph XVI

that rebuts the presumption that the scope of the right remains unchanged.

       A. The common-law right against compelled self-incrimination arose
from a limited context, but evolved considerably from the time of the founding
of the Republic to Georgia’s constitutionalization of the doctrine in 1877.

      The right against compelled self-incrimination borrowed heavily from

England’s common law of criminal procedure. See Charles Gardner Geyh, The

Testimonial Component of the Right Against Self-Incrimination, 

, 620 (1987). The English crown’s deprivation of the colonists’ “rights

as Englishmen,” including the right against compelled self-incrimination by

conducting judicial proceedings that were “very inquisitional and ofttimes

overbearing,” was a contributing cause of the Revolutionary War. See R. Carter

Pittman, The Colonial and Constitutional History of the Privilege Against Self-

Incrimination in America, 

, 783-787 (1935). And in the midst

of that conflict, several states were quick to adopt constitutions that protected

                                       23
this right, and the Fifth Amendment became part of the United States

Constitution in 1791. Id. at 764-765, 787; see also Geyh, 36 Cath. U. L. Rev. at

620; Leonard W. Levy, Origins of the Fifth Amendment: The Right Against

Self-Incrimination 409 (1968).

      Despite the importance placed on this essential right during the founding

era, the scope of the right historically is not entirely clear and has been the

subject of considerable debate. See, e.g., John Fabian Witt, Making the Fifth:

The Constitutionalization of American Self-Incrimination Doctrine, 1791-1903,

, 832 (1999) (“Scholars have had considerable difficulty

explaining the original meaning of the American constitutional self-

incrimination provisions — both state and federal.”). But it is clear that the right

had limited value at common law and in colonial America, particularly given

that defendants were not allowed to testify under oath and usually were

unrepresented. The right against compelled self-incrimination took on greater

significance as more defendants gained access to lawyers to defend them. And

as states in the new Republic began to allow defendants to testify under oath in

their own defense if they wished to do so, states also afforded suspects

protection from adverse comment by prosecutors if defendants declined the

                                        24
option. Thus, between the time of the founding and the time Georgia adopted the

1877 Provision, the right against compelled self-incrimination took on greater

relevance.

      (i) Given the nature of criminal procedure in pre-Revolution England,
the common-law right arose in a limited context, finding its roots in objections
to compelled self-incrimination through torture or required oaths.

      The conventional view (which long ago found its way into our case law)

is that the right against compelled self-incrimination is modeled on the maxims

nemo tenetur seipsum accusare and nemo tenetur seipsum prodere (either of

which might be translated, “no man is bound to accuse himself”). See Calhoun,

; State v. Davis, 

 (Or. 2011). Under the

nemo tenetur principle, which was believed to originate in the medieval canon

law of the Roman Catholic Church, one was not obliged to confess punishable

offenses to judges and prosecutors (although one had a spiritual obligation to

reveal all sins at confession as a condition of absolution). See United States v.

Gecas, 

 (11th Cir. 1997). But this right against public

confessions ceased to apply once a formal accusation was made, at which point

it was incumbent upon the accused to prove his innocence. See 

;

Akhil Reed Amar & Renee B. Lettow, First Amendment First Principles: The

                                       25
Self-Incrimination Clause, 

, 896 (1995). This right became

more established centuries later in response to the Star Chamber and

ecclesiastical Court of High Commission, which required suspects to swear in

advance to answer truthfully questions about their religious and political beliefs

— forcing subjects either to lie under oath (and thereby risk eternal damnation)

or refuse to take the oath (and thereby risk being whipped or pilloried). Doe v.

United States, 

 (

, 101 LE2d 184) (1988);

Andresen v. Maryland, 

 (

, 49 LE2d 627) (1976).

In response to these harsh inquisitorial practices, Puritans claimed the benefit

of the nemo tenetur principle and refused to take the oath or to testify against

themselves. See Gecas, 

; John H. Langbein, The Historical

Origins of the Privilege Against Self-Incrimination at Common Law, 

, 1073 (1994).

      Common-law courts at some point began to apply this principle as giving

rise to a right that “no man [was] bound to incriminate himself” on “any charge”

or “in any court[.]” See Davis, 

 n.2 (citing John Henry

Wigmore, 4 Treatise on the System of Evidence in Trials at Common Law §

2250 (1905)). And so a defendant in colonial times was protected from being

                                       26
forced to be his own accuser — although the protection ceased once an

accusation was made — and such a defendant had a right against compelled self-

incrimination, as long as he claimed the right. Levy, supra, at 375 (“Historically,

it has been a fighting right: unless invoked, it offered no protection. It vested an

option to refuse answer but did not bar interrogation nor taint a voluntary

confession as improper evidence.”). The right against compelled self-

incrimination thus was an important limitation on the use of coercion. See Eben

Moglen, Taking the Fifth: Reconsidering the Origins of the Constitutional

Privilege Against Self-Incrimination, 

, 1100-1101 (1994).

     (ii) Outside of the context of torture or compulsory oaths, the right
against compelled self-incrimination was of limited value at common law in
England and the American colonies.

      Nonetheless, any right against compelled self-incrimination had minimal

value in the criminal justice systems of pre-colonial Britain and the American

colonies, given pretrial investigatory procedures, defendants’                legal

incompetence to testify under oath, and defendants’ inability to rely on counsel

to make their case for them. See Moglen, 92 Mich. L. Rev. at 1092-1093, 1104.

Early colonial pretrial procedure was based on British statutes that required a

defendant to be examined after apprehension. See, e.g., Amar & Lettow, 93

                                        27
Mich. L. Rev. at 897. When an accusation was made, a magistrate or justice of

the peace would conduct a preliminary examination of the defendant, but not

under oath; the defendant’s statements and admissions were used at a later

criminal trial; and if he refused to answer questions, his refusal could be

admitted for whatever inferences might be drawn from his refusal. See Amar &

Lettow, 93 Mich. L. Rev. at 897-898; Langbein, 92 Mich. L. Rev. at 1059-1062.

This practice continued in England without modification until 1848, when

Parliament provided that the accused must be advised that he need not answer

and warned that what he said might be used against him. See Levy, supra, at

375; Langbein, 92 Mich. L. Rev. at 1061.

     Defendants’ incompetence to testify under oath — as well as their need

to defend themselves without the aid of lawyers — made questions about the

breadth of the right against compelled self-incrimination somewhat

“metaphysical” as a matter of English common law through the time of the

founding of America. See De Luna v. United States, 308 F2d 140, 150 (5th Cir.

1962). English law deemed the accused incompetent to testify under oath until

1898. Id. And until defendants were permitted to have counsel represent them

at trial, defendants had little choice but to speak — unsworn — in their own

                                     28
defense. See Gecas, 

; Langbein, 92 Mich. L. Rev. at 1049-1054.

This often resulted in an inquisitorial trial proceeding whereby the accused was

bullied into answering questions. If he refused, he subjected himself to the

possibility that the judge would tell the jury that the accused would have cleared

himself if he were innocent. See Mitchell v. United States, 

, 333-

334 (

, 143 LE2d 424) (1999) (Scalia, J., dissenting); State v.

Bentley, 54 S2d 137, 139 (La. 1951). In England, defense lawyers were first

permitted in ordinary felony trials in the 1730s, but as a practical matter were

used only infrequently until the 1780s. Albert W. Alschuler, A Peculiar

Privilege in Historical Perspective: The Right to Remain Silent, 

, 2655 (1966).

      The practice of magistrates questioning accused perpetrators before trial

continued in our new Republic, with suspects’ silence continuing to be

introduced against them at trial at least until the mid-1800s. See Mitchell, 

 (Scalia, J., dissenting); Amar & Lettow, 93 Mich. L. Rev. at

897-898. One exception was that the prisoner’s statement could not be used

against him if he were first placed under oath, perhaps reflecting the view that

questioning under oath was unduly coercive, if not equivalent to torture. See

                                       29
Twining v. New Jersey, 

 (

, 53 LE 97) (1908)

(“[C]ommitting magistrates were authorized to take the examination of persons

suspected, which if not under oath, was admissible against him on his trial[.]”).

      As in England, American criminal defendants in the years after the

founding could not testify under oath but were nonetheless forced to defend

themselves at trial without counsel. It was only in 1859, beginning with Maine,

that American states began permitting criminal defendants to give sworn

testimony, with every state except Georgia allowing such testimony by the end

of the century. See Ferguson v. Georgia, 

 (

, 5

LE2d 783) (1961). Very few defendants were represented by counsel at the

framing of the Bill of Rights and the early state constitutions. See Alschuler, 94

Mich. L. Rev. at 2656; Levy, supra, at 369; compare id. at 376 (“In the seven

colonies where the right to counsel was provided even in felony cases, the right

against self-incrimination was well secured in common-law trials.”).7

       (iii) The right against compelled self-incrimination took on increased
value — and greater scope — in the early years of the Republic, as the criminal
justice system changed.



      7
          Georgia was not one of those colonies. Levy, supra, at 376.

                                             30
       Although English criminal practices, as well as the English common law’s

narrow view of the right against compelled self-incrimination, carried over into

the American colonies, American law and practice in this regard evolved

considerably in the decades leading up to Georgia’s adoption of a constitutional

provision against compelled self-incrimination in 1877. For one thing, the nature

of the privileges that comprised the common-law right against self-incrimination

— the witness privilege rule, the confession rule, the rule barring defendants

from testifying under oath, and the rule disqualifying interested parties from

testifying8 — changed over the course of the nineteenth century. See Witt, 77

Tex. L. Rev. at 838-906. In particular, the rule barring criminal defendants from

testifying under oath eventually was abolished. See Ferguson, 

;

see also De Luna, 308 F2d at 150-151; Amar & Lettow, 93 Mich. L. Rev. at 891

& n.153.




       8
        The witness privilege protected a witness from being compelled by court process to
make disclosures that might incriminate him or subject him to civil liability or moral
disgrace. The confession rule excluded the admission of pretrial confessions extracted under
oath or by force, threat, or promise. The rule disqualifying parties prevented criminal
defendants from testifying under oath at all. And the rule disqualifying witnesses for interest
disqualified witnesses based on financial interest in the outcome of the case. See Witt, 77
Tex. L. Rev. at 834.

                                              31
      Abolition of the rule against criminal defendants testifying under oath

came with a parallel trend. Most states, at the same time they granted defendants

the ability to testify, prohibited comment by the prosecution on the accused’s

failure to do so. See, e.g., Ferguson, 

 (“The majority of the

competency statutes of the States forbid comment by the prosecution on the

failure of the accused to testify, and provide that no presumption of guilt should

arise from his failure to take the stand.”). Although, as discussed further below

in Division IV (C) (ii), Georgia was an outlier and did not deem defendants

competent to testify under oath until 1962, see Ga. L. 1962, p. 133, § 2, this

Court’s case law reflects an understanding, prior to the end of the nineteenth

century, that adverse inferences may not be drawn from a defendant’s failure to

make an (unsworn) statement at trial. See, e.g., Bird v. State, 

(7) (1874), overruled on other grounds by Shelton v. State, 

(

) (1965).

      In addition, the increased availability of defense counsel strengthened the

right against compelled self-incrimination, as defendants could remain silent and

let counsel defend them. Langbein, 92 Mich. L. Rev. at 1068-1071; Amar &

Lettow, 93 Mich. L. Rev. at 897. It is against this backdrop — in which a

                                       32
longstanding right took on new importance in changing circumstances — that

Georgia adopted the 1877 Provision.

      B. The scope of the right against compelled self-incrimination
originally covered more than oral testimony.

      As discussed above, some states and the federal government

constitutionalized the right against compelled self-incrimination soon after the

Revolution. The state constitutions contained bills or declarations of rights that

phrased their clauses in terms of a right not to be compelled to “give evidence”

or to “furnish evidence.” See United States v. Hubbell, 

 (

, 147 LE2d 24) (2000) (Thomas, J., concurring) (outlining

development of state and federal constitutional rights against compelled self-

incrimination after the Revolution). Although the Fifth Amendment used the

phrase “to be a witness” instead of the language “to give evidence” or “to

furnish evidence” that was proposed by several states advocating for a bill of

rights,9 these textual differences did not necessarily reflect a difference in

meaning. See 

, 53-54 (citing legal sources and dictionaries published

at the time of the founding that defined “witness” as a “person who gives or

      9
        The federal constitution guarantees that no person “shall . . . be compelled in any
criminal case to be a witness against himself.” U. S. Const. amend. V.

                                            33
furnishes evidence,” and noting that this broad definition is consistent with the

meaning given to the “same term (albeit in plural form) in the Sixth

Amendment’s Compulsory Process Clause”). Georgia’s constitutional clause

uses the term “testimony,” but this semantic change also has been viewed as not

altering the scope of a right rooted in centuries of history.10 See Olevik, 

; cf. United States v. Burr, 

 (D. Vir. 1807)

(Marshall, C. J.) (equating “testimony” with “evidence”: “The propriety of

introducing any paper into a case, as testimony, must depend on the character

of the paper[.]”). These textual differences historically have been understood as



       10
         Even prior to the adoption of the 1877 Provision, Georgia provided by statute that
no one “charged” in a criminal proceeding shall be “compellable to give evidence for or
against himself or herself.” See Code of 1867 § 3798 (2). Similar language is found in our
current Code. See OCGA § 24-5-506 (a). We have construed a prior iteration of the statute
as applying “only to those who have been charged with an offense — i.e., accused in a
returned or proposed charging document — at the time they are called to testify.” State v.
Lampl, 

 (2) (

) (2015) (construing former OCGA § 24-9-20
(a)).
       In the same section, the 1867 Code also included a subsection that was not limited to
those who were “charged,” providing, “No person shall be compellable to answer any
question tending to criminate himself or herself.” Code of 1867 § 3798 (3). Although the
1877 constitutional provision in some respects tracked this statutory provision, it differed in
a key respect, in that it prohibited compelled incriminating “testimony,” which is broader
than the statutory reference to being compelled “to answer any question.” The statutory
provision later was amended to mirror the constitutional language, see Code of 1895 § 1011
(3), and was carried forward into a subsequent code, see Code of 1933 § 417. Presumably
given that it was superfluous of the constitutional provision, it did not survive the adoption
of the Code of 1981.

                                              34
not reflecting a difference in meaning because they all refer to the same common

law. See Twining, 

 (“[A]ll the states of the Union have, from time

to time, with varying form but uniform meaning, included the privilege in their

Constitutions, except the States of New Jersey and Iowa, and in those states it

is held to be part of the existing law.”); Counselman v. Hitchcock, 

 (

, 35 LE 1110) (1892) (reviewing federal provision and

various state provisions and stating that “however differently worded, [the

provisions] should have as far as possible the same interpretation”); Wigmore,

supra, § 2263 (“It is therefore immaterial that the witness is protected by one

Constitution from ‘testifying,’ or by another from ‘furnishing evidence,’ or by

another from ‘giving evidence,’ or by still another from ‘being a witness.’ These

various phrasings have a common conception, in respect to the form of the

protected disclosure.”).11 And, as we discuss below, around the time the Georgia

       11
         In Olevik, we said that Day did not explain its “broad interpretation” of the self-
incrimination provision, noting that the plain meaning of the word “testimony,” as that term
was defined around the time the provision was first adopted in 1877, would not have covered
affirmative acts. Olevik, 

 (2) (c) (ii) n.6. Our statement in Olevik should not
be misunderstood as endorsing a view that terms are to be construed in isolation; such a
reading would ignore other statements in Olevik in which we explained that we are to
construe the text of a constitutional provision in the light of the broader legal context
surrounding the adoption of that provision. See 

 (2) (c) (i) (the meaning and effect
of a constitutional provision “is to be determined in connection, not only with the common
law and the constitution, but also with reference to other statutes and the decisions of the

                                             35
Constitution of 1877 was ratified, a prominent view was that the right against

compelled self-incrimination protected the accused from being compelled to

provide, do, or say anything that might tend to incriminate him. See Geyh, 36

Cath. U. L. Rev. at 621-622.

      (i) Our precedent prior to passage of the 1877 Constitution described
the scope of the right as protecting “any evidence.”

       Prior to the adoption of a provision on self-incrimination in the 1877

Constitution, Georgia courts described the right against compelled self-

incrimination as forbidding a “man . . . to accuse himself of any crime, or to

furnish any evidence to convict himself of any crime[.]” Marshall v. Riley, 

 (1849) (emphasis in original); see also Higdon v. Heard, 

 (1853) (“The doctrine is well settled in England and America, that no man

is bound to accuse himself of any crime; or to furnish any evidence to convict

himself of any crime.”). In Marshall, this Court held that the right against

compelled self-incrimination, based on the common law and “considered as one

of our constitutional rights” (despite there being no such express state

constitutional right at that time), applied in civil proceedings and protected a


courts” (citation and punctuation omitted)). We did not consider the constructions of other
jurisdictions’ self-incrimination provisions in Olevik, as we do here.

                                            36
person from being compelled to answer interrogatories that would subject him

to a penalty or forfeiture or punishment for a crime, “or have a tendency

thereto.” Marshall, 

 (syllabus).12 Just two years after this right

received express protection in our state constitution, this Court construed that

provision as barring a defendant from being “compelled to [in]criminate himself

by acts or words.” Day, 63 Ga. at 668 (2) (syllabus).

       Day’s language mirrors language used in our decisions issued prior to the

1877 Constitution, and the temporal proximity of these decisions to adoption of

the 1877 Constitution make them critical indicators of the original public

meaning of the self-incrimination protections of that constitution. See District

of Columbia v. Heller, 

 (

, 171 LE2d 637)

(2008) (“[T]he examination of a variety of legal and other sources to determine

the public understanding of a legal text in the period after its enactment or

ratification . . . is a critical tool of constitutional interpretation.” (emphasis in

       12
          Marshall left open the question whether the right could protect a person from
answering questions where he legally could not face charges (due to the alleged expiration
of the statute of limitations in that case). Id. at 372-373. This Court in Higdon partly
answered that question, holding that civil litigants could be compelled to answer questions,
but such answers could not be used in a criminal prosecution against the individual, in order
to comport with the United States Constitution (which we applied even though the Self-
Incrimination Clause of the Fifth Amendment had not yet been incorporated against the states
— indeed, before there was a Fourteenth Amendment). Higdon, 14 Ga. at 258-259.

                                             37
original)); Noel Canning v. N.L.R.B., 

 (D.C. Cir. 2013) (“The

interpretation of the Clause in the years immediately following the

Constitution’s ratification is the most instructive historical analysis in discerning

the original meaning. Indeed, such early interpretation is a critical tool of

constitutional interpretation because it reflects the public understanding of the

text in the period after its ratification.” (citing Heller, 

)

(punctuation omitted) aff’d on other grounds, N.L.R.B. v. Noel Canning, 

 (

, 189 LE2d 538) (2014); Ezekiel v. Dixon, 

 (1847) (“[G]reat regard ought to be paid to the interpretation put upon a

statute by the sages of the law who presided at the time, or shortly after its

passage[.]”).

      (ii) Around the time Day was decided, the United States Supreme Court
construed the Fifth Amendment as covering more than oral testimony.

      In 1886, less than ten years after passage of the 1877 Constitution, the

United States Supreme Court unanimously held that compelling production of

a defendant’s private papers to be used as evidence against him was equivalent

to compelling him to be a witness against himself in violation of the Fifth

Amendment. See Boyd v. United States, 

 (

, 29


                                         38
LE 746) (1886) (“a compulsory production of the private books and papers of

the owner of goods sought to be forfeited . . . is compelling him to be a witness

against himself, within the meaning of the Fifth Amendment [of] the

Constitution”); see also Wigmore, supra, § 2264 (“it is universally conceded”

that one may refuse production of one’s documents by asserting the right against

compelled self-incrimination). “[T]he Court linked its interpretation of the Fifth

Amendment to the common-law understanding of the self-incrimination

privilege.” Hubbell, 

 (Thomas, J., concurring) (citing Boyd,

).

      Boyd was about the compelled production of existing documents, and said

nothing about compelling a person to act. The United States Supreme Court later

rejected the view that the Fifth Amendment protected affirmative acts, see Holt

v. United States, 

 (

, 54 LE 1021) (1910), but for

many years did not distinguish between spoken words and physical evidence

reflecting written words. A shift in Fifth Amendment jurisprudence began to

occur with the United States Supreme Court’s decisions in Schmerber v.

California, 

 (

, 16 LE2d 908) (1966), where

the Court held that the Fifth Amendment protects the accused only from being

                                       39
compelled to testify against himself or to give evidence “of a testimonial or

communicative nature,” and a series of cases in the 1970s that effectively

overruled Boyd’s rationale.13 See Amar & Lettow, supra, 93 Mich. L. Rev. at

885-888 (describing Schmerber as the turning point in the Court’s Fifth

Amendment jurisprudence with respect to a distinction between words and

physical evidence). Before this jurisprudential shift, for almost 80 years, the

United States Supreme Court had construed the Fifth Amendment right against

compelled self-incrimination to protect an individual from compulsory

incrimination through his testimony or personal records. See Bellis v. United

States, 

 (

, 40 LE2d 678) (1974) (citing cases

predating Schmerber that “reflect[ed] the Court’s consistent view that the

privilege against compulsory self-incrimination should be limited to its historic

function of protecting the natural individual from compulsory incrimination

through his own testimony or personal records” (citation and punctuation



       13
         See Andresen, 

 (holding that defendant’s Fifth Amendment
right was not violated by introduction of business records because the statements were
“voluntarily committed to writing” before police arrived and were seized pursuant to a valid
search warrant); Fisher v. United States, 

 (

, 48 LE2d 39)
(1976) (holding that government could compel defendant to produce incriminating papers
and Fifth Amendment protected only “testimonial” aspects of the transfer).

                                            40
omitted)); see also Hubbell, 

 (Thomas, J., concurring) (noting that

the United States Supreme Court had “not always taken the approach to the Fifth

Amendment that [it] follow[s] today”).

      (iii) Other states had interpreted their self-incrimination clauses to
protect affirmative acts.

      If the construction of the Fifth Amendment today is narrower than it was

in the last quarter of the nineteenth century, today’s construction of many states’

constitutional provisions has narrowed even further — they now largely mirror

the Supreme Court’s present interpretation of the Fifth Amendment, but began

more broadly than the Fifth Amendment was ever interpreted. Despite the

State’s claim here that Georgia’s broad self-incrimination protection is a unique

outlier in this country and in other countries that have followed English common

law, that same rule was quite ordinary a century ago. See Amar & Lettow, 93

Mich. L. Rev. at 884 (“[M]any courts in the late nineteenth and early twentieth

centuries held that using the defendant’s body as physical evidence was in effect

compelling the defendant to be a witness against himself.”); Geyh, 36 Cath. U.

L. Rev. at 621-622 (“[A] majority of courts addressing the issue were of the

opinion that the right to refuse to give evidence, to be a witness, or to testify .


                                        41
. . meant that the accused cannot be compelled to do or say anything” that was

incriminating (citation and punctuation omitted)).

      From the middle of the nineteenth century through the middle of the

twentieth century, at least nine states, including three of our neighboring states

(Alabama, North Carolina, and Tennessee), concluded that defendants could not

be compelled to perform affirmative acts that were incriminating; many of these

jurisdictions cited their state constitutional provision or the common law. See,

e.g., Allen v. State, 

 (Md. 1944) (state constitution protected the

accused from being compelled to perform an evidence-producing act, such as

requiring the accused, while on the witness stand, to try on a hat that had been

worn by the culprit); Beachem v. State, 

 (Tex. Crim. App.

1942) (forcing the defendant to speak for identification purposes violated state

constitution); Ward v. State, 

 (Okla. Crim. App. 1924)

(forcing the defendant to make “experiments and comparisons” in open court in

presence of jury violated state constitution); State v. Height, 

,

938-939 (Iowa 1902) (relying on cases applying self-incrimination privilege to

bar results from compulsory physical exam); Cooper v. State, 6 S 110, 111

(Ala. 1888) (Alabama constitution made it “unlawful to force the witness to give

                                        42
(or make) evidence against himself,” and state could not accomplish the same

result indirectly by having witness testify that defendant refused to make

footprints for comparative analysis); Nolen v. State, 

 (Tex. Ct.

App. 1883) (recognizing that compelled acts of defendants are as inadmissible

as compelled confessions); People v. Mead, 

 (Mich. 1883)

(explaining that a criminal defendant cannot be compelled to try on a shoe to

determine whether tracks found at the scene of the crime were his); Stokes v.

State, 

 (1875) (holding that forcing defendant to make a footprint

for the jury was illegally compelled “testimony”); State v. Jacobs, 

,

260 (N.C. 1858) (reversing conviction where trial court forced defendant to

stand up at his trial because the “generous spirit of the common law” protected

accused from being compelled to furnish evidence against himself).14

       Forty years after recognizing that affirmative acts were protected under the

state’s constitutional protection against compelled self-incrimination, the

       14
           Compare State v. Garrett, 71 NC 85, 88 (1874) (holding that coroner’s testimony
of the incriminating evidence that he saw when the defendant was forced to unwrap her hand
did not violate defendant’s self-incrimination right, and distinguishing Jacobs on the basis
that, in that case, it was the prisoner who was compelled to exhibit himself to the jury). We
recognize that some of these decisions involved compulsion at trial, but this does not negate
the fact that the courts construed their state’s self-incrimination law to cover affirmative acts.
Moreover, the State in this case is asking us to do away with the affirmative act standard
altogether, not simply limit it to a trial setting.

                                               43
Oklahoma Criminal Court of Appeals held that the results of involuntary

chemical testing in a DUI case would be inadmissible as violative of the state’s

constitutional self-incrimination clause that was “broad enough to encompass

more than just oral testimony,” and ruled that the court properly instructed the

jury not to consider the results of the test if it found that the defendant did not

voluntarily submit to such testing. See Cox v. State, 

(Okla. Crim. App. 1964). Several years later, Utah reached a similar conclusion

that its state’s constitutional ban against compelled self-incrimination covered

affirmative acts. See Hansen v. Owens, 

 (Utah 1980) (forcing

defendant to furnish a handwriting exemplar violated state constitutional self-

incrimination provision).

      (iv) The meaning of our constitutional right against compelled self-
incrimination does not change simply because that of other jurisdictions has
changed.

      In the light of the history described above, the State’s sweeping

pronouncements that Day and Calhoun were uniquely out of step with the rest

of the country are simply wrong. And though many jurisdictions have since

abandoned their affirmative acts standard in favor of a right that is limited to the




                                        44
present scope of the Fifth Amendment,15 the decisions of sister states cannot

change the meaning of a preexisting provision of the Georgia Constitution. One

court that altered course did so without any meaningful analysis. Hill v. State,

366 S2d 318, 322 (Ala. 1979) (noting that it had abandoned its prior

interpretation because the Alabama privilege “offers the same guarantee as that

contained in the Federal Constitution” and “the interpretation of the [F]ifth

[A]mendment in Schmerber is likewise controlling as to the interpretation of art.

I, section 6 of the Alabama Constitution.”). Other courts did attempt to

determine the original meaning of their self-incrimination protections, relying

in part on Wigmore’s treatise on the issue and the United States Supreme

Court’s opinion in Schmerber, but there is no indication that any of them shared

the same legal circumstances relevant here — a constitutional provision that had

received a consistent and definitive construction and whose text was carried

forward into a new constitution after that consistent construction. See, e.g.,

American Fork City v. Crosgrove, 

 (Utah 1985) (rejecting

affirmative act standard and limiting scope of self-incrimination right to


      15
         One scholar has posited that limiting protection became the majority view largely
as a “practical consideration” due to an ever-increasing reliance on physical evidence as a
form of proof. See generally Geyh, 36 Cath. U. L. Rev. at 623-628.

                                            45
“evidence of a testimonial or communicative nature”); State v. Thomason, 

 (Okla. Crim. App. 1975) (holding that state constitutional

right protects against only “testimonial compulsion or the equivalent thereof”);

Olson v. State, 

 (Tx. Crim. App. 1969) (overruling

Beachem after noting that its case law had been inconsistent, as some opinions

sanctioned compelled affirmative acts, and that Beachem had been “seriously

eroded” by subsequent cases).

      And although the United States Supreme Court has said that the Fifth

Amendment does not apply to affirmative acts, see, e.g., United States v. Wade,

 (

, 18 LE2d 1149) (1967); Holt, 

-253, in reaching these holdings the Supreme Court of course did not have

to confront even the stare decisis implications of contrary prior holdings, much

less a consistent and definitive prior construction of language then essentially

replicated in the constitution being applied. The situation we face is different,

and so these decisions from the United States Supreme Court and other state

courts shed little light on Paragraph XVI’s meaning.

      Similarly, we reject the State’s argument that our jurisprudence is at odds

with foreign jurisdictions that have adopted English common law and we should

                                       46
therefore conform to the prevailing view. Of course, we may consider other

jurisdictions’ definition of a right that is rooted in English common law, but

only to the extent those jurisdictions consider the original public meaning of

those rights based on shared language, history, and context. It is the role of this

Court, not the United States Supreme Court, other states’ courts, or courts of

foreign countries, to construe the meaning of the Georgia Constitution in the

light of its particular language, history, and context.

     C. Olevik properly applied the consistent and definitive construction
presumption.

      The State argues that we were wrong in Olevik to give Paragraph XVI of

the 1983 Constitution the same meaning that we had given to the materially

identical clause in the Constitutions of 1877, 1945, and 1976. First, the State

argues that we misapplied the prior construction canon referenced in Olevik

because the canon applies only to ambiguous provisions, which the State argues

Paragraph XVI is not. We rejected this argument above in Division II (B) and

need not address it further. Second, the State argues that we should have

interpreted Paragraph XVI according to its plain, literal meaning divorced from

our prior precedent on the right against compelled self-incrimination, because


                                        47
the Constitution of 1983 was meant as a “new” constitution that should be

interpreted on its own terms. We reject the State’s arguments. Olevik rightly

applied the presumption that flowed from this consistent and definitive

construction, because, as is explained below, there is no evidence that Paragraph

XVI of the 1983 Constitution was meant to depart from the settled construction.

      (i)   Our case law shows a consistent and definitive construction.

      Our case law shows that at least six of our decisions under the 1877

Constitution cited or applied the holding of Day, all of them reinforcing the

construction that the self-incrimination clause protected affirmative acts, but did

not prohibit taking evidence from the defendant. See Calhoun, 

(illegal seizure of items from the defendant’s person did not constitute

compelled production of incriminating evidence); Elder v. State, 

 (1) (

) (1915) (ordering a new trial where jury heard evidence

that defendant was ordered by sheriff to put foot in track); Dozier v. State, 

 (1) (

) (1899) (evidence that sheriff found pistol in

defendant’s pocket was admissible); Evans v. State, 

 (

)

(1899) (reversing for new trial where defendant compelled to put hand in

pocket); Franklin v. State, 

 (3) (1882) (“inducing” defendant

                                        48
to take off his shoes and socks did not violate right, as fact he wore decedent’s

shoes, and not any act he did under coercion, was the incriminating evidence);

Blackwell v. State, 

 (1) (1881) (right violated where defendant

was required to stand up at trial so that a witness could verify that the

defendant’s leg had been amputated in a way that corresponded to tracks left at

the crime scene).16

       Calhoun more clearly explained the rationale behind Day that the

constitutional right was “as broad as . . . the common-law privilege from which

it is derived.” 

. We repeatedly relied on Calhoun for the principle

that evidence taken from a defendant, as opposed to being produced by the

defendant, does not violate the self-incrimination clause. See, e.g., McIntyre v.

State, 

 (1) (

) (1940); Herndon v. State, 

          In Drake v. State, 

 (1885), decided after Day, we said that the
constitutional provision “means that, when a person is sworn as a witness in a case, he shall
not be compelled to testify to facts that may tend to criminate him.” 

 (2). This
statement, though sounding like a holding, was not the actual holding of the case, because
the Drake Court held that taking clothing from a defendant and submitting that clothing to
the jury would not violate the defendant’s constitutional right against compelled self-
incrimination. 

 (2). This holding is consistent with the principle, laid out in
subsequent cases, that the right against compelled self-incrimination is not violated where
a defendant is compelled only to be present so that certain incriminating evidence may be
procured from him. See Olevik, 

 (2) (c) (iii) (citing cases). Moreover, no
subsequent decisions followed the reasoning of Drake by limiting Paragraph XVI to oral
testimony.

                                             49
832, 844 (3) (

) (1934);17 Hysler v. State, 

, 409 (1) (

) (1918); Groce v. State, 

 (1) (

) (1918).

      Our opinion in Aldrich, 

, reaffirmed that the 1945

Constitution, which contained the same language as the 1877 Constitution,

carried forward the principle enunciated in Day and its progeny; we cited or

distinguished Aldrich in at least three other cases decided prior to the adoption

of the 1976 Constitution. See Strong v. State, 

 (

)

(1973) (extraction of blood did not violate right against compelled self-

incrimination under Georgia or United States Constitutions), overruled in part

as stated in Olevik, 

 (2) (a) & n.2 (noting that Williams v.

State, 

 (

) (2015), overruled Strong’s search and

seizure holding but not its holding regarding Paragraph XVI); Creamer v. State,

 (3) (

) (1972) (removal of bullet from

defendant did not violate Georgia right against compelled self-incrimination);

Dennis v. State, 

 (2) (

) (1970) (requirement that

driver drive onto scales was not “compulsion” because there was no criminal



      17
        In Herndon, we acknowledged that we had rejected prior requests to overrule
Calhoun. See Herndon, 

 (3).

                                        50
penalty for refusing). There were also several cases decided by this Court under

the Constitution of 1976 that continued Day’s and Calhoun’s interpretation of

the scope of the right. See Raines v. White, 

 (

)

(1981); Strickland v. State, 

 (18) (

) (1981);

Fouts v. State, 

 (3) (

) (1977).

      In short, the compelled self-incrimination provision at issue, which has

remained materially the same since the 1877 Constitution, has received a

consistent and definitive construction from its inception through the ratification

of the 1983 Constitution. And

      [a]t no point through this history was the constitutional language
      changed to abrogate Day’s interpretation, nor did we reconsider
      Day. To the contrary, we have consistently and repeatedly applied
      the state constitutional protection against compelled
      self-incrimination in accord with Day.

Olevik, 302 Ga. at 240 (2) (c) (ii) (emphasis added).

      Indeed, even before Olevik, we continued to apply the Day/Calhoun

construction of the self-incrimination right in the 1983 Constitution. Before our

decision in Olevik, we referred repeatedly to the broader right afforded by

Paragraph XVI compared to the right against compelled self-incrimination under

the Fifth Amendment. See Bell v. State, 

 (3) n.4 (

382) (2013) (“[T]he Georgia provision has been construed liberally to limit the

State from forcing an individual to affirmatively produce any ‘evidence, oral or

real,’ regardless of whether or not it is testimonial.” (citing Muhammad v. State,

 (2) n.1 (

) (2007)). And these were not just mere

restatements of the law. We actually applied the Day/Calhoun affirmative act

standard to Paragraph XVI, and in one instance found that the right had been

violated. See, e.g., Simpson v. State, 

 (3) (

)

(2011) (requiring a defendant to turn over his clothes to the police for inspection

did not violate the right against compelled self-incrimination); Quarterman v.

State, 

 (4) (

) (2007) (statute requiring certain

convicted felons to provide a DNA sample did not violate self-incrimination

right because it forced the felon “only to submit his or her body for the purpose

of having the evidence removed”); Brown v. State, 

 (10) (

) (1993) (requiring defendant to provide handwriting exemplar

violated the right); Green v. State, 

 (2) (

)

(1990) (collection of urine — a substance naturally excreted by the human body

— did not violate the right); Batton v. State, 

 (3) (

) (1990) (removal of defendant’s shoes for evidentiary purposes did not

                                        52
violate right because his “presence [was] all the cooperation required”); Ingram

v. State, 

 (7) (

) (1984) (requiring defendant to

strip to his waist and be photographed did not compel self-incrimination);18

State v. Thornton, 

 (2) (

) (1984) (the taking of

dental impressions does not violate right), overruled on other grounds by Neal

v. State, 

 (

) (2012) (Hunstein, C. J., concurring,

joined by all Justices).

        Nothing in the cases decided under the 1983 Constitution reflects a

considered departure from the consistent and definitive construction that had

been placed on the self-incrimination right. In Olevik, we simply applied the

presumption that flowed from the consistent and definitive construction existing

prior to the 1983 Constitution, and concluded that the presumption regarding the

construction of Paragraph XVI — that it covers affirmative acts — was not

rebutted by a clear indication to the contrary. Olevik, 

 (2) (c)

(ii).

        The State argues that around the time of the adoption of the 1983



        18
         The holding of Ingram was recently applied in Anglin v. State, 

(4) (

) (2017) to reach the same result.

                                          53
Constitution, one commentator and one court asserted that Georgia had “shown

signs of retreating” from the affirmative act standard. See American Fork City,

 (citing Note, The Georgia Right Against Self-Incrimination:

Historical Anomaly or Vanguard of Justice? 

, 1109-1113

(1981)). But none of the case law cited by the Utah court and the law student’s

note on which it relies — Presnell v. State, 

 (

) (1978),

reversed on other grounds by Presnell v. Georgia, 

 (

, 58

LE2d 207) (1978); Fouts, 

; and English v. State, 

 (

) (1975) — actually demonstrates a clear departure from Day and its

progeny.

      In Presnell, we rejected the defendant’s claim that testimony of an expert

witness should not have been allowed because it was based upon a compelled

psychiatric exam, whereby the defendant was forced to be a witness against

himself as to his mental capacity. Presnell, 

 (7). We cited both

Creamer and cases decided under the Fifth Amendment in rejecting that claim.

But a careful reading shows that the defendant did not complain about

performing any incriminating act; instead, he argued only that his statements

were compelled. Presnell, 

 (7). Because only statements were at

                                       54
issue, it is unsurprising that the Presnell Court would analyze the state and

federal claims similarly, and the Court’s citation to Creamer was merely to

support the conclusion that certain evidence can be taken from the defendant

without offending the Georgia Constitution. 

 (7) (noting that

defendant did not claim that any incriminating statements made were admitted

at trial, and concluding that his statements to a medical examiner were not

testimonial in nature).19 Thus, Presnell does not signal a retreat from Day.

      In Fouts, the defendant sought to suppress the introduction of samples of

his hair, which he claimed was in violation of his right against compelled self-

incrimination. 240 Ga. at 43-44 (3). Fouts did not explain which constitutional

right — federal or state — was at issue, but the opinion noted that the defendant

had been given Miranda warnings and primarily relied on federal case law

regarding testimonial protections, suggesting that only the Fifth Amendment

right was at issue. Id. at 43-44 (3) (citing Creamer only as a “see also”).

Similarly, in English, the Court did not identify whether the federal or state self-

incrimination right was at issue, and simply cited to Creamer and Strong to

summarily conclude that removal of traces of substances from a defendant’s

      19
           Whether we would reach the same outcome in Presnell today is not before us.

                                            55
hands was constitutional because it was “non-testimonial evidence.” 

 (3) (b). The cursory analysis in both cases cannot be taken as a reflection

that we were abandoning the longstanding affirmative act standard in favor of

the “testimonial” standard of the Fifth Amendment. Moreover, the removal of

evidence from the defendants in Fouts and English would not have qualified as

compelled affirmative acts protected by Georgia’s constitutional right against

compelled self-incrimination. Thus, these cases do not negate the consistent and

definitive construction that had existed up through the adoption of the 1983

Constitution.

      (ii) Nothing in the adoption of the 1983 Constitution rebuts the
consistent and definitive construction of Day and its progeny.

      The State argues that, in the absence of evidence that the public

understood the full scope of the provision that became Paragraph XVI and the

lack of evidence that the voters accepted or rejected the given meaning of the

former provision when ratifying the 1983 Constitution, we should presume that

the voters understood the right against compelled self-incrimination as applying

only to oral testimony. But contrary to the State’s argument, it is “the

understanding of the text by reasonable people familiar with its legal context”


                                      56
that is important, not whether every citizen understood the particular meanings

of a constitutional provision. See Jerman, 

 n.1 (Scalia, J.,

concurring in part and concurring in the judgment) (“[W]here . . . many federal

courts have read [certain text in a given] way . . . it could be said that such

uniform and longstanding judicial interpretation had established the public

meaning of the text, whether the Members of Congress were aware of the cases

or not. That would be the understanding of the text by reasonable people

familiar with its legal context.”); Baptiste, 

 (Nahmias, J.,

concurring) (explaining that contemporaneous sources are useful in seeking to

interpret the meaning of a constitutional text “because they demonstrate what

intelligent and informed people at the time understood the language . . . to

mean”). Day and its century-long progeny were sufficiently consistent and

definitive such that the legal context of the text of Paragraph XVI extended to

affirmative acts.

      The State also argues that we erred in Olevik because we ignored the

“realities that surrounded the ratification of” the 1983 Constitution. First, the

State argues that the 1983 Constitution was ratified not as an amendment to

existing organic law, but as an entirely new constitution that should be read

                                       57
anew according to its plain language. In support, the State cites to a provision

that repealed “all previous Constitutions and amendments thereto,” see Ga.

Const. 1983, Art. XI, Sec. I, Par. VI, and to various statements — some made

near the time of ratification and some made in March 2018 — that the 1983

Constitution was meant as a “new and revitalized organic law” that should

“stand on its own terms.”20 The fact that certain individuals labeled the 1983

Constitution as “new,” or even “revitalized,” does not rebut the presumption

arising from Day’s consistent and definitive construction. We had characterized

the 1945 Constitution as “new,” see Wheeler v. Bd. of Trustees of Fargo

Consolidated School Dist., 

 (2) (

) (1946), and yet

determined the meaning of certain provisions of that constitution by applying

the prior constructions of similarly worded provisions found in the 1877

Constitution, see, e.g., Aldrich, 

 (determining meaning of

self-incrimination clause and citing cases interpreting and decided under 1877

Constitution); Hancock, 

 (construction of provision of 1877

       20
           In March 2018, the State, apparently for purposes of this appeal, interviewed two
of the many people instrumental in creating the 1983 Constitution and reported that they
opined that the 1983 Constitution was meant to be “new” and was to “stand” or “be looked
at” on its own terms. As we discussed in Olevik, this sort of subjective legislative intent has
little relevance, especially for a constitution ratified by the voters. Olevik, 

-
238 (2) (c) (i).

                                              58
Constitution regarding purposes for which counties may levy taxes was

controlling as to meaning of same provision in 1945 Constitution). To presume

that the meaning of text changes simply because it is readopted in a “new”

constitution ignores the heavy emphasis courts place on history and context in

interpreting text.

      And even for those who may believe that the subjective views of

individual drafters are relevant, there is no evidence whatsoever that the “new”

1983 Constitution was meant to wipe away nearly 200 years of Georgia

constitutional law. To the contrary, the transcripts from the various meetings of

the committee charged with revising Article I (our Bill of Rights) showed that

the committee meant to keep existing law — even if all of the implications of

the language were not fully explored or even understood — unless otherwise

stated. See, e.g., Select Committee on Constitutional Revisions, 1977-1981

(“Select Committee”), Transcript of Meetings, Committee to Revise Article I,

meeting of the Subcommittee to Revise Section I, Oct. 4, 1979, pp. 3-4 (Justice

Jesse Bowles, subcommittee chairman, noting that Georgia’s Bill of Rights had

remained “pretty much” the same since the first constitution and proposing that

the committee review the 1976 Constitution “paragraph by paragraph” and

                                       59
revise as necessary). Moreover, there was acknowledgment that many of the

rights enshrined in the Bill of Rights came from the English common law. 

meeting of the Subcommittee on Origin and Structure of Government, Oct. 26,

1979, p. 54.

      Some committee members also expressly admitted that they did not

understand the meaning of certain clauses of the Bill of Rights. The committee

nevertheless elected not to change the words of such clauses, because to do so

would signal that a change was intended. 

 meeting of Subcommittee to

Revise Section I, Oct. 4, 1979, p. 69 (noting that the search and seizure clause

had “been construed so many times” and a “tremendous body of law” developed

on the words of that clause that “if we change much of that we’re going to open

a complete new field”); id., p. 97 (noting that the committee would “open up a

keg of worms” if it “monkey[ed] with” the double jeopardy clause); id., pp. 103-

106 (notwithstanding members’ uncertainty about meaning of phrase

“corruption of blood,” Justice Bowles noted that the phrase had been defined in

case law, and another committee member suggested the phrase remain in the

light of that case law); id., meeting of Subcommittee on Rights of Persons, Oct.

25, 1979, p. 51 (Justice Bowles noted “change should be made where change is

                                       60
necessary but” courts view a change in words as “an intention on the part of the

framers to give it a different meaning from the meaning that theretofore

existed”); id., Nov. 9, 1979, meeting of Full Committee, pp. 22-29 (one

committee member proposed removing the word “remonstrance” from provision

on right to assemble and petition, but majority of committee voted to keep the

provision as written after argument was made that the alternative language

omitting the word would narrow the right). This evidence shows that, rather than

starting with a clean slate, the framers of the 1983 Constitution generally

understood themselves to be proposing the preservation of existing law when

they left language unchanged.

      In conclusion, none of the State’s arguments discussed above undermines

the soundness of Olevik. Although we do not determine conclusively that Day

was correctly decided, that case (and others like it) established a well-settled

interpretation of the self-incrimination right, one that was not unique but within

the mainstream of American judicial decisions at the time. That interpretation

was carried forward into subsequent Georgia Constitutions, and there was no

error in Olevik in concluding that Paragraph XVI preserved the protections of

prior constitutional self-incrimination provisions because the previous

                                       61
provisions were materially identical.21 Thus, Olevik rightly decided that

Paragraph XVI covers state-administered breath tests. Now that we have

concluded that Olevik should stand, we must address the consequence of Olevik

that this case presents: does Paragraph XVI prohibit the State from admitting

into evidence Elliott’s exercise of her constitutional right to refuse to submit to

a breath test, as OCGA §§ 40-5-67.1 (b) and 40-6-392 (d) say the State may do?

     IV. Admission of evidence that a defendant refused to submit to a
chemical test of breath pursuant to OCGA § 40-5-67.1 (b) violates Paragraph
XVI of the Georgia Constitution.

      Our reaffirmance of Olevik means that Elliott’s refusal to submit to a

breath test fell within the protections of Paragraph XVI. But that does not

necessarily mean that introduction of evidence of that refusal at trial violates

Paragraph XVI. That issue presents a separate question. After thorough review

of the relevant history and case law, we conclude that the admission of such a

refusal violates Paragraph XVI.

      The pre-Revolution English common-law right against compelled self-


      21
          Because we conclude that Olevik was correctly decided, it is unnecessary to
consider whether we should retain that decision under the doctrine of stare decisis. See
Kimble v. Marvel Entertainment, LLC, ___ U. S. ___, ___ (

, 192 LE2d 463)
(2015) (“stare decisis has consequence only to the extent it sustains incorrect decisions;
correct judgments have no need for that principle to prop them up”).

                                           62
incrimination did not preclude admission of a defendant’s refusal to incriminate

herself or adverse inferences therefrom. And the text of Paragraph XVI does not

by its plain terms preclude admission of evidence that a defendant refused to

speak or act. But legal developments in the century between the Revolution and

our state’s adoption of the 1877 Provision — in particular, decisions of this

Court preceding and contemporaneous with that adoption — support the

conclusion that the 1877 Provision prohibited admission of a defendant’s

invocation of his privilege against compelled self-incrimination. We find

nothing sufficient to rebut the presumption that the 1983 Constitution

incorporated this meaning and thus conclude that Paragraph XVI generally

prohibits admission of a defendant’s pretrial refusal to speak or act. And we

conclude that OCGA §§ 40-5-67.1 (b) and 40-6-392 (d) violate the Georgia

Constitution by allowing the admission of a defendant’s refusal to submit to a

breath test to prove that the defendant had been drinking alcohol.

      A. The United States Supreme Court’s interpretation of the Fifth
Amendment is unpersuasive as to the meaning of the Georgia Constitution on
this point.

      To be clear, this conclusion is not dictated by the United States Supreme

Court’s decision in Griffin v. California, 

 (

, 14 LE2d

                                      63
106) (1965), in which that Court held that the Fifth Amendment barred state

prosecutors from commenting on a defendant’s failure to testify. To the extent

that such a decision was rooted in text, history, and context shared by the 1877

Provision, it would be persuasive as we determine the original public meaning

of that provision. But Griffin was not so rooted; indeed, it has been met with

considerable criticism since the day it was decided on the ground that it “lacks

foundation in the [federal] Constitution’s text, history, or logic.” Mitchell, 

 (Thomas, J., dissenting); see also, e.g., 

 (Scalia,

J., dissenting, joined by Rehnquist, C. J., and O’Connor and Thomas, JJ.); Carter

v. Kentucky, 

 (

, 67 LE2d 241) (1981)

(Powell, J., concurring); Griffin, 380 U. S. at 617-623 (Stewart, J., joined by

White, J., dissenting). To the extent the Fifth Amendment and Paragraph XVI

(and its predecessors) share some textual features, Griffin is unhelpful to our

task because it contains no textual analysis. And to the extent that it considered

the context and history of the Fifth Amendment (which it does not appear to

have done, either), it certainly is not definitive as we consider the history and

context of the 1877 Provision, because, as discussed below, the history and

context of the Fifth Amendment is different from the history and context of the

                                       64
1877 Provision. Thus, Griffin cannot answer the question before us today, and

we must undertake our own analysis.

       B. Although the text of the 1877 Provision considered in isolation does
not clearly preclude admission of evidence that a defendant refused a breath
test, that does not end our analysis of the provision’s meaning.

       Although the state constitutional provision at issue here is sometimes

referred to as embodying a right against self-incrimination, by its text the 1877

Provision prohibited only “compelled” self-incrimination. Paragraph XVI also

uses the term “compelled.” “Compel” is generally defined as to “constrain” or

“force” a person to do something. See Webster’s New World Dictionary 289 (2d

College ed. 1980).22 Just as asking a person to take a breath test, without more,

is not compelling that person to do any act, admission of evidence of a

defendant’s refusal to submit to a breath test does not fit within a natural

understanding of the word “compel.” Rather, such a refusal merely is a

consequence of a DUI arrestee’s choice between two options. Both options are

decidedly unpalatable given that both likely would result in incriminating

evidence: agree to the breath test and risk an incriminating result, or refuse and


       22
          At the time of the 1877 Constitution, the term “compel” similarly was defined as
“[t]o drive or urge irresistibly” or “[t]o take by force.” Noah Webster, A Dictionary of the
English Language 80 (1878) (emphases added).

                                            65
face driver’s license suspension and the risk that a jury will assume the arrestee

had been drinking. But this poor option set is merely a consequence of there

being probable cause to arrest a person for driving under the influence. And

making a choice between two unpalatable options is still a choice. See

McGautha v. California, 

 (

, 28 LE2d 711) (1971)

(“The criminal process, like the rest of the legal system, is replete with situations

requiring the making of difficult judgments as to which course to follow.

Although a defendant may have a right, even of constitutional dimensions, to

follow whichever course he chooses, the Constitution does not by that token

always forbid requiring him to choose.” (citation and punctuation omitted));

State v. Modlin, 

, 620 (Neb. 2015) (citing cases from several

jurisdictions for proposition that although the legal consequences of an

administrative license revocation make refusal of a chemical test a difficult

choice to make, “the difficulty of such choice does not render consent

involuntary”), cited in Kendrick v. State, 

 (

)

(2016).

      But as we explained in Division II, we do not read the text in isolation;

rather, “[w]e interpret a constitutional provision according to the original public

                                         66
meaning of its text,” Olevik, 

 (2) (c) (i), which requires

considering its context. Where, as here, a constitutional provision incorporates

a pre-existing right, the provision cannot be said to create that right — it merely

secures and protects it. See Nunn v. State, 

 (1846) (relying on

constructions of other states’ constitutional provisions securing right to bear

arms in construing Second Amendment, because “these instruments confer no

new rights on the people which did not belong to them before” (emphasis in

original)). And where the right enshrined in the constitution was one found at

common law, that constitutional right is understood with reference to the

common law, absent some clear textual indication to the contrary. See Virginia

v. Moore, 

 (

, 170 LE2d 559) (2008) (“We look

to the statutes and common law of the founding era to determine the norms that

the Fourth Amendment was meant to preserve.”); United States v. Wong Kim

Ark, 

 (

, 42 LE 890) (1898) (“The language of the

Constitution, as has been well said, could not be understood without reference

to the common law.”); Central of Ga. R. Co., 

 (“In construing a

constitution, a safe rule is to give its words such significance as they have at

common law; especially if there is nothing in the instrument to indicate an

                                        67
intention by its framers that the language in question should have a different

construction.”). We thus proceed to consider pre-Revolution English common

law on the right against compelled self-incrimination, then subsequent

developments that provide important context for the original public meaning of

the 1877 Provision.

      C. The historical record prior to Georgia’s adoption of the 1877
Provision indicates the provision prohibited admission of a defendant’s refusal
to incriminate herself or the drawing of adverse inferences therefrom.

      The pre-Revolution English common-law right did not preclude admission

of a defendant’s refusal to incriminate herself or adverse inferences therefrom.

But the second half of the nineteenth century was a time of great change in the

American criminal justice system. And decisions of this Court in the years

leading up to, and around the time of, the adoption of the 1877 Provision

indicate a significant deviation from the common law of England on the

question of whether a defendant’s silence could be presented as evidence of

guilt. We thus conclude that the original public meaning of the 1877 Provision

prohibited admission of a defendant’s refusal to incriminate herself or the

drawing of adverse inferences therefrom.

      (i)   Inferences from refusal to testify were permissibly drawn at common

                                       68
law.

       The common law as it was understood in 1776 does not support

construing the 1877 Provision as forbidding admission of a defendant’s refusal

to incriminate herself or the drawing of adverse inferences therefrom. The

common law as it was understood in 1776 did not prohibit a trial court from

admitting evidence that a defendant refused to speak or otherwise provide

incriminating evidence against himself. And it did not prohibit jurors from

drawing adverse inferences from such a refusal. Until 1848, English magistrates

would conduct preliminary examinations of defendants without advising them

that they need not answer or that their unsworn statements could be used against

them in a later trial. See Langbein, 92 Mich. L. Rev. at 1060-1061. If the

defendant refused to answer questions, his refusal could be admitted, and any

adverse inferences might be drawn therefrom. See id. at 1061-1062. The

question of whether adverse inferences could be drawn from the accused’s

failure to “testify” at trial did not even arise, as defendants were deemed

incompetent to testify under English law until 1898. See De Luna, 308 F2d at

150. Moreover, defendants’ inability to rely on a lawyer to speak for them meant

that a failure to speak amounted to no defense at all. See Langbein, 92 Mich. L.

                                       69
Rev. at 1049-1053. And if a defendant did fail to answer questions, he risked the

jury, with the judge’s blessing, drawing an inference that his silence was

indicative of guilt. See Mitchell, 

 (Scalia, J., dissenting);

Bentley, 54 S2d at 139; Alschuler, 94 Mich. L. Rev. at 2631. Thus, had the 1877

Provision afforded constitutional protection to the common-law right as it was

understood in 1776, that provision would not have precluded the admission of

— or an instruction that the jury may draw adverse inferences from — a

defendant’s failure to testify. But that is not what the 1877 Constitution did.

       (ii) Developments between the time of Georgia’s adoption of the
common law of England and our ratification of the 1877 Constitution indicate
that the 1877 Provision barred admission of a defendant’s refusal to speak or
act.

      By the end of the nineteenth century, American jurisdictions generally

experienced a shift in the admissibility of defendants’ refusal to incriminate

themselves. The practice of magistrates questioning the accused before trial,

with suspects’ silence being used against them at trial, continued in America

into the 1800s. See Mitchell, 

 (Scalia, J., dissenting); Amar

& Lettow, 93 Mich. L. Rev. at 897-898. And to the extent that defendants

declined to make a statement at trial, the practice of drawing an adverse


                                       70
inference persisted into the 1800s as a general matter. See Alschuler, 94 Mich.

L. Rev. at 2631. But by the end of the nineteenth century every state but Georgia

had given defendants the ability to testify under oath. See Ferguson, 

. And at the same time they granted defendants competency to testify,

most states prohibited comment by the prosecution on the accused’s failure to

do so. 

; Wigmore, supra, § 2272 (“In the United States, almost

universal legislation has decreed, in varying phraseology, that the inference shall

not be availed of.”); see also, e.g., State v. Williams, 

 (S.D.

1898); State v. Pearce, 

 (Minn. 1894); State v. Baldoser, 

 (Iowa 1893); Staples v. State, 

 (Tenn. 1890); Hunt v.

State, 

 (Tex. App. 1889); Long v. State, 

(1877).23

       Georgia in at least one sense was an outlier, prohibiting defendants from

testifying under oath in their own defense at trial until 1962. See Ga. L. 1962,


       23
          Perhaps more consistently with English common law, it also was fairly clear in
American courts by the latter half of the nineteenth century that no adverse inferences could
be drawn from the silence of a witness who was not a criminal defendant. See Greenleaf on
Evidence, at § 451 (11th ed. 1863) (“If the witness declines answering, no inference of the
truth of the fact is permitted to be drawn from that circumstance.”); see also People v.
Maunausau, 

 (Mich. 1886); Phelin v. Kenderdine, 

(1853); Rose v. Blakemore, 7 Geo. IV 382, 383-384 (1826).

                                             71
pp. 133, 135, § 2. In the meantime, defendants could make only an unsworn

“statement” on their own behalf. See Code of 1873 § 4637 (as amended by Ga.

L. 1874, pp. 22, 22-23, § I) (“In all criminal trials in this State, the prisoner shall

have the right to make, to the Court and Jury, such statement in the case as he

or she may deem proper, in his or her defense, said statement not to be under

oath, and to have such force only as the jury may think right to give it: Provided,

the prisoner shall not be compelled to answer any questions on cross-

examination, should he or she think proper to decline to answer such

questions.”).24 And it was not until defendants were afforded competency to

testify that the legislature explicitly prohibited commenting on a defendant’s



       24
          A similar rule applied to pretrial questioning by magistrates. An 1860 statute
provided that the magistrate “shall always permit the defendant to make his own statement
of the transaction (not under oath), if he desires to do so” and that “[w]henever such
statement is made, it shall be the duty of the court to reduce it to writing, and return it with
the other papers to the Superior Court, in the event of a commitment.” Code of 1860 § 4614.
But the statute did not specify whether an accused’s refusal to answer questions could be
admitted at trial. Id. An 1819 Georgia justice of the peace manual indicates that an accused’s
answers to pretrial questioning by a magistrate should be admitted at trial, as long as his
answers were not given under oath or extracted by threats or promises, but, again, did not
explicitly address refusals to speak. See Augustin S. Clayton, The Office and Duty of a
Justice of the Peace: and a Guide to Clerks, Constables, Coroners and Other Civil Officers
According to the Laws of the State of Georgia 92, 122, 127 (Milledgeville: S. Grantland,
1824) (providing that if an accused had been sworn before being examined, his answers
could not be used against him at trial — “because that is making a party a witness against
himself” — and that confessions induced by “threats or promises” were inadmissible as
well).

                                              72
failure to speak in his own defense, providing that “[t]he failure of a defendant

to testify shall create no presumption against him, and no comment shall be

made because of such failure.” Ga. L. 1962, pp. 133, 135, § 1.25

       Nonetheless, decisions of this Court prior to the adoption of the 1877

Provision, consistent with those of other jurisdictions at the time, indicate that

it was impermissible to draw adverse inferences from a defendant’s failure to

make such a statement. In 1874, this Court overturned a criminal conviction

when the jury was charged that it might consider the defendant’s refusal to make

a statement at trial. See Bird, 

 (7). This decision appeared to reflect

the Court’s view that allowing a defendant to make a statement, but forbidding

him from testifying, did defendants more harm than good. The Court noted it did

“not think that the statute giving this right to a defendant intended that it should

be counted against him, if he did not avail himself of it,” and the authoring

Justice observed that he had “never known or heard of but one instance where

it supposed that the right had availed anything” and thus it would be “terrible”

if a defendant’s failure to speak could be considered against him. Bird, 

         This statutory provision, with slight alteration, can now be found in OCGA § 24-5-
506 (b), which provides: “The failure of an accused to testify shall create no presumption
against the accused, and no comment shall be made because of such failure.”

                                            73
at 589 (7). And two years later — the year before Georgia adopted the 1877

Constitution — we reversed a conviction because the trial court had allowed the

prosecutor to argue that the defendant’s objection to being required to produce

certain papers amounted to an admission of guilt. See Russell v. State, 

 (4) (1876). We reasoned that the trial court had improperly

endorsed, rather than corrected, the State’s improper argument that an assertion

of rights through counsel amounted to an admission of guilt, when in fact that

assertion “was no evidence of his guilt whatever[.]” 

 (4). In neither

case did we cite the Fifth Amendment or any other statutory or constitutional

provision, any court decision, or really any authority at all, and the State argues

that Bird applies a mere common-law evidentiary rule and cannot say anything

about Georgia’s constitutional provision given that Bird predated it. It may well

be correct that Bird and Russell were common-law cases. But the common law

as it was understood in 1877 informs the scope of the 1877 Provision.

      Consistent with that approach, other cases decided prior to the adoption

of the 1877 Provision afforded protection to civil parties or third-party witnesses

faced with questions the answers to which might implicate them criminally. See

Pinkard v. State, 

 (2) (1860) (trial court properly told State’s

                                        74
witness he need not answer question about certain conversation if answer could

incriminate him; court “should have ruled out the whole of the conversation”);

Higdon, 

 (gaming law requiring discovery in civil case not

unconstitutional under Fifth Amendment because answers in such cases could

not be used against the witness in a criminal case).

      We also decided similar cases shortly after the adoption of the 1877

Provision. We said early on that the 1877 Provision meant a witness did not

have to answer questions that might subject the witness to criminal prosecution,

or even ones that merely “tend[ed] to degrade” the witness. See Gravett v. State,

 (6) (1884) (relying on state constitutional provision to hold

that trial court did not err by ruling that State’s witness, who had testified that

she had a child, did not have to answer whether she had ever been married). Less

than a decade after the adoption of the 1877 Provision, citing both the Georgia

and federal constitutions, we found no error in excluding depositions offered in

a civil case to show that the original plaintiff (since deceased) had refused to

answer questions the answers to which might implicate him in the crime of

doing business on a Sunday. See Harrison v. Powers, 

 (6), (8)

(1886). We wondered “[o]f what worth would the protection [of the right

                                        75
against self-incrimination] be if a party availing himself of it incurred detriment

or loss by so doing.” 

 (8). And 30 years later, we again held that a

witness’s refusal to answer questions pursuant to the right found in the state

constitution could not be admitted as evidence in a subsequent proceeding, even

in a civil case. See Loewenherz v. Merchants’ and Mechanics’ Bank of

Columbus, 

 (

) (1916) (court in civil real estate dispute

erred in admitting rule nisi and answer in contempt case, which evidenced

party’s failure to answer questions before grand jury). We quoted that same

language from Harrison, as well as an Iowa Supreme Court decision to the effect

that “‘[i]t would indeed be strange if the law should confer upon a witness this

right as a privilege, and at the same time should permit the fact of his availing

himself of it to be shown as a circumstance against him.’” 

 at 560 (quoting

State v. Bailey, 

 (Iowa 1880)). One of our sister state courts

relied on our decision in Loewenherz, as well as decisions from other

jurisdictions, to conclude that it had “become well settled by the great weight of

authority that no inference of the existence of the incriminating fact is permitted

to be drawn from the witness’ claim of privilege, and that therefore the witness’

claim of privilege is not a proper matter of evidence for the jury’s

                                        76
consideration.” State v. Weber, 

 (Mo. 1917) (citing

Loewenherz). These decisions, coming after the adoption of the 1877 Provision,

could not change its original public meaning. But, as previously discussed, their

temporal proximity to its adoption makes them good indicators of its meaning.

       The State argues that, to the extent the Georgia constitutional provision

incorporates the Bird and Loewenherz lines of cases, such incorporation would

retain a common-law limitation to refusals to testify in formal court proceedings.

But those cases are not so limited. Harrison involved a refusal to answer

questions in a deposition, not a formal court proceeding. 

 (6).26

And we have never limited Paragraph XVI to compelled self-incrimination in


       26
           Citing Simpson v. Simpson, 

 (

) (1974), the State suggests
that we implicitly have overruled Harrison. In Simpson, a custody case, we held that there
was no error in drawing a negative inference from the refusal of a party and third-party
witness to answer questions about their allegedly illicit relationship. The State argues that
Simpson implicitly overruled Harrison to the extent that Harrison applied a no-adverse-
inference rule to a refusal to testify outside of a formal court proceeding. But Simpson made
no such distinction. And to the extent Simpson impliedly overruled Harrison, it did so only
to the extent that Harrison suggested that an adverse inference from one’s refusal to testify
in a civil case could not be drawn in that same civil case. See Simpson, 

 (“We
think it is clear that no inference of guilt can be drawn from a privileged refusal to testify in
a criminal case nor can the exercise of the privilege in a civil case be used in a subsequent
criminal case against the party. However, these cited Georgia cases do not hold that it is
impermissible to draw an unfavorable inference in a civil case from the privileged refusal to
testify in that case. There is considerable authority that such an inference can be drawn in
civil cases.”). This makes sense — where no criminal proceeding is at issue, it is particularly
difficult to argue that the State is involved in compelling a person to incriminate himself. At
any rate, if we did overrule Harrison, we did so only as to circumstances not applicable here.

                                               77
the context of a court proceeding. As discussed below, we have specifically

applied Paragraph XVI to bar a criminal prosecution that was based on a refusal

to provide incriminating evidence by the side of the road. See Aldrich, 

.

       Similarly helpful is a line of cases, beginning just twelve years after the

adoption of the 1877 Provision, in which this Court deemed improper negative

prosecutorial comment on the defendant’s failure to make a statement at trial.

Robinson v. State, 

 (9) (

) (1889). Although we

concluded that the error was rendered harmless by the court’s jury instruction,

we held that “counsel for the State should not argue to the jury from the

omission of the accused to make a statement[.]” 

 (syllabus). This

became an oft-cited rule — at least until the United States Supreme Court in

Griffin held that the Fifth Amendment barred such comments even in state court

trials. See, e.g., O’Dell v. State, 

 (4) (

) (1904); Minor

v. State, 

 (1) (

) (1904); Caesar v. State, 

 (

) (1906); Barker v. State, 

 (

) (1907); Thornton v.

State, 

 (1) (

) (1940); Parks v. State, 

 (1) (

) (1951); Jordan v. State, 

 (3) (92

                                        

) (1956).27 Neither Robinson nor its progeny rely upon precursors to

Paragraph XVI; rather, they appear to be common-law cases. But this suggests

that the common law they apply was part of the preexisting common law that

was incorporated into the 1877 Provision. The historical record in the years

leading up to and shortly after Georgia adopted the 1877 Provision lead us to

conclude that it prohibited admission of a defendant’s refusal to speak or act as

evidence against him.

     D. No subsequent developments clearly altered the meaning of the
1877 Provision.

       Of course, the 1877 Provision does not apply today; Paragraph XVI of the

1983 Constitution does. As explained in Division II, a constitutional provision

— like Paragraph XVI — that is retained from a previous constitution without

material change is strongly presumed to retain the original public meaning that

provision had at the time it was first adopted. Thus, we presume that Paragraph

XVI as it is found in the 1983 Constitution carries the same meaning as that of

the 1877 Provision. Although not ironclad, this presumption is strong. As also


       27
         Where the defendant did make a statement at trial, the prosecution was permitted
to note for the jury that the defendant’s statement had failed to deny some material fact
established by the state’s evidence. See, e.g., Howard v. State, 

 (2) (

) (1952); Tolbert v. State, 

 (

) (1913).

                                             79
explained in Division II, a constitutional clause that is readopted into a new

constitution and that has received a consistent and definitive construction is

presumed to carry the same meaning as that consistent construction. But unlike

the question at issue in Olevik, this second presumption does not arise regarding

refusals. Although there is ample case law from which we conclude that

Paragraph XVI precludes admission of a defendant’s refusal to speak or act and

the drawing of adverse inferences therefrom, much of that case law relates to the

common law informing the original public meaning of the 1877 Provision; few

of the cases were actual interpretations of the 1877 Provision of the sort

necessary to support the presumption arising from a consistent and definitive

construction.

      In 1964, well before Georgia readopted the provision against compelled

self-incrimination as part of the 1976 and 1983 Constitutions, we held that the

provision precluded a conviction based on a defendant’s failure to do an

incriminating act — in that case, his refusal to drive his truck onto scales. See

Aldrich, 

. Although the applicable statute in that case went further

than the one at issue here, by making refusal itself a crime, see 

, our

holding was consistent with an understanding that the constitutional provision

                                       80
prohibited using refusal to support a criminal prosecution.28

       The State relies heavily on a line of cases — since overruled at least as a

matter of federal law — that allowed a jury to hear about and be charged on a

defendant’s silence in the face of statements made in his presence that

implicated him in criminal activity, essentially as an exception to the hearsay

rule. Dating back to the nineteenth century, Georgia evidence law provided that

silence when the circumstances require an answer, denial, or other conduct may

amount to an admission. See Code of 1860 § 3713; see also Cobb v. State, 

 (5) (1859) (upholding conviction, on the basis that the jury

should hear all relevant evidence, against a challenge that the jury should not

have heard that the defendant refused to confess when confronted by a jailer

with a letter from a co-defendant). For a long time, we applied this tacit

admission provision with little consideration of whether the circumstances might

implicate either a state or federal right against compelled self-incrimination,



       28
         A later case distinguishing Aldrich in finding no violation of the constitutional right
in a conviction under a new version of the statute whereby a driver’s refusal to drive his
vehicle onto scales could result in a loss of license is not at odds with that understanding,
given that the conviction at issue in that case was not based on the defendant’s refusal to
comply with a direction to drive onto the scales, but on the evidence obtained when he
voluntarily complied. See Dennis v. State, 

 (1) (

) (1970).

                                              81
even when law enforcement was involved in the interaction that gave rise to the

inference. See, e.g., Smiley v. State, 

 (

) (1923)

(rejecting challenge to witness’s testimony that when she implicated the

defendant in the crime in the presence of a deputy sheriff and the defendant, the

defendant remained silent); see also Bennett v. State, 

 (2)

(

) (1973); Bloodworth v. State, 

 (3) (

) (1961); Kalb v. State, 

 (2) (

) (1943);

McCormick v. State, 

 (1) (

) (1932).

      At one point we rejected, albeit with little analysis, a self-incrimination

challenge to the tacit admission rule. Emmett v. State, 

 (

)

(1943), involved the admissibility of testimony that the defendant failed to

respond to certain statements made by a co-defendant in his presence, as well

as the propriety of a jury instruction that the defendant’s failure to deny the

statements would amount to an admission of the statements by the defendant if

the circumstances were such that “an answer or denial was required[.]” 

 (2). The defendant claimed that because the co-defendant’s

incriminating statements were made while the defendant was in jail, in the

presence of a sheriff and prosecutor, and charged with assault, he had a right to

                                       82
remain silent such that the instruction was contrary to the Fifth and Fourteenth

Amendments and the 1877 Provision. 

 (2). We concluded that “in our

opinion” the fact that the accused was in custody at the time of his silence did

not render evidence of that silence inadmissible and that the instruction was not

erroneous. 

 (2).

      In at least one instance prior to the Miranda decision in 1966, however, we

questioned whether the tacit admission rule was always compatible with the

right against compelled self-incrimination. See Johnson v. State, 

(

) (1921) (plurality opinion) (even if evidence at issue — a deputy’s

testimony that, while in jail charged with homicide, defendant did not comment

upon accomplice’s statements that arguably implicated her in the crime — were

admissible, it was not “positive evidence of guilt” sufficient to support a guilty

verdict). Ultimately, decades after the adoption of the 1945 Constitution, we at

least partially overruled our tacit admission case law as irreconcilable with

Miranda. See Howard v. State, 

 (

) (1976)

(stating that Bennett, Kalb, and Emmett must be overruled in part). In the light

of Miranda, we construed the tacit admission statute as not authorizing a charge

based on silence or acquiescence by a person when they are in police custody,

                                       83
on the theory that police interrogation is not a circumstance requiring an answer

or denial. 

.29

       We did not do away with the tacit admission rule altogether. See Emmett

v. State, 

 (1) (

) (1979) (applying tacit

admission rule and distinguishing Howard as involving custodial interrogation).

Under our current Evidence Code, the tacit admission rule is embodied in

OCGA § 24-8-801 (d) (2) (B), which provides that among the party-opponent

admissions excluded by the hearsay rule is a “statement of which the party has

manifested an adoption or belief in its truth[.]” But whatever we have said about

the constitutionality of the tacit admission rule, this line of cases is not enough

to rebut the presumption that Paragraph XVI precludes admission of a

defendant’s refusal to incriminate himself to the same extent that the 1877

Provision did. This line of cases did not hold that a defendant’s silence was

       29
          Miranda, of course, is a construction of the Fifth Amendment — not the Georgia
Constitution. And it’s not obvious that Howard overruled Emmett’s holding as to the Georgia
Constitution. Although Howard quoted the Georgia Constitution, the only articulated bases
in Howard for overruling Emmett were decisions by the United States Supreme Court
regarding the meaning of the United States Constitution. See Howard, 

;
see also State v. Herrera-Bustamante, 

 (

) (2018) (“Howard’s
holding was based explicitly on Miranda and other federal constitutional case law[.]”). This
is a meager basis on which to conclude that we also adopted a new construction of the
Georgia Constitution. In concluding that Howard does not require affirmance here, we need
not and do not decide whether it overruled Emmett as a matter of state constitutional law.

                                            84
always admissible, only that it was admissible when circumstances dictated that

the defendant should have responded, rendering the silence an adoptive

admission. See Emmett, 

 (2) (rejecting defendant’s argument that

his custodial status at the time his co-defendant made certain statements

rendered his failure to deny those statements inadmissible, and that it was

improper to instruct the jury that such silence could amount to an admission, but

providing that “in so doing the judge should be careful to inform the jury that

they are to determine whether under all the circumstances an answer or denial

was required”); Graham v. State, 

 (

) (1903).

And the State does not argue that Elliott’s refusal to submit to a breath test

would constitute an adoptive admission excluded by the hearsay rule under our

current Evidence Code.

      The State argues that the United States Supreme Court in South Dakota

v. Neville, 

 (

, 74 LE2d 748) (1983), held that

admission of a defendant’s refusal to consent to a blood-alcohol test does not

violate the Fifth Amendment, suggesting that this means we cannot conclude

that refusal to submit to a breath test is “coerced” and thus that admission of that

refusal is barred by Paragraph XVI. 

. But Neville’s holding followed

                                        85
from the Court’s prior holding that undergoing a blood-alcohol test is not “of a

testimonial or communicative nature” and thus is not protected by the Fifth

Amendment. Schmerber, 

. The Court reasoned in Neville

that “the values behind the Fifth Amendment are not hindered” when the state

offers the suspect a choice between submitting to a blood-alcohol test that the

state could legitimately compel or having his refusal used against him. Neville,

. But our decision does not turn on the meaning of the word

“coerced” as a textual matter, nor on “the values behind the Fifth Amendment,”

but, rather, on the meaning of the common-law right against self-incrimination

when it was first enshrined into the Georgia Constitution in 1877 and then

readopted in our current Constitution. Neville’s understanding of what

constitutes “coercion” thus is not relevant to the ultimate question before us

here.

      E.    Because Elliott had the right to refuse to submit to a breath test
under Paragraph XVI, admission of evidence of her refusal violates that state
constitutional right against compelled self-incrimination.

        The State posits a number of additional arguments as to why evidence of

Elliott’s refusal to submit to a breath test should be admissible in its criminal

prosecution of her. For one, the State contends that DUI cases are

                                       86
“extraordinary” in that suspects are deemed to have waived their right to refuse

a breath test in exchange for the privilege of driving on the roads of this state.

But we have clearly held that the fact that an officer reads a suspect the implied

consent notice and otherwise complies with implied consent procedures does not

mean that the suspect gives actual and voluntary consent to a particular test for

Fourth Amendment purposes. See Williams v. State, 

 (

) (2015). Similarly, a person’s mere decision to drive on the roads of

Georgia does not amount to actual agreement that a refusal to consent to a breath

test may be admissible in her criminal prosecution.

      The State also contends that, under a strict scrutiny framework, it may use

evidence of Elliott’s refusal because in encroaching on any fundamental right

against self-incrimination, the State is acting on a compelling interest in

effectively prosecuting drunk drivers in a way that is narrowly tailored to serve

that interest. But this argument would allow similar exceptions for all serious

crimes; surely the State’s interest in prosecuting murderers is at least as strong

as its interest in prosecuting drunk drivers. The State has failed to identify any

case in which we have applied strict scrutiny to Paragraph XVI or any other

constitutional rule of criminal procedure. Strict scrutiny is a concept generally

                                        87
invoked when statutes are challenged under the First Amendment or on equal

protection or substantive due process grounds. See generally Richard H. Fallon,

Jr., Strict Judicial Scrutiny, 

, 1268-1269 (2007). Elliott’s

challenge to the statute is not based on First Amendment or equal protection

grounds. And although she raises due process concerns, her primary argument

rests on Paragraph XVI. Where a specific constitutional amendment provides

“an explicit textual source of constitutional protection against a particular sort

of government behavior,” the express provision, “not the more generalized

notion of substantive due process,” controls the analysis. Albright v. Oliver, 

 (

, 127 LE2d 114) (1994) (plurality opinion) (citation

and punctuation omitted).

      The State also argues that a holding that Paragraph XVI prohibits

admission of a defendant’s refusal to submit to a breath test would contravene

Olevik’s holding that the implied consent warning is not per se coercive, as well

as our subsequent decisions in Fazio v. State, 

 (

)

(2017), and Schmitz v. State, 

 (

) (2017), relying on

Olevik to reject challenges to the implied consent notice. To the extent the State

argues that our holding here is inconsistent with and nullifies Olevik, the State

                                       88
fails to appreciate the differences between the claims decided in Olevik and

those raised here. In Olevik, we resolved the defendant’s claim that he was

coerced into submitting to a breath test because the language of the implied

consent notice did not sufficiently inform him that he was entitled to refuse.

Given that argument, the consequence of refusing was not among the particular

aspects of that language the defendant challenged. We rejected his argument and

held that the implied consent notice sufficiently informed a defendant of the

right to refuse such that the notice was not per se coercive. Olevik, 

 (3) (a). Here, we do not address whether Elliott was coerced into

submitting to a breath test under Paragraph XVI, but what consequences flow

from her assertion of the right to refuse to submit to the test. We recognize that

our holding here may affect a totality-of-the-circumstances inquiry into whether

a defendant voluntarily submitted to a breath test where the State first threatened

that, if she refused, that would be evidence against her at trial. But that question

is not presently before us.

      The State’s remaining arguments, citing state statutes, general principles

of state criminal law, and policy concerns, fail to appreciate the constitutional

nature of the right at issue. We acknowledge that the State has a considerable

                                        89
interest in prosecuting DUI offenses (and thereby deterring others), and that our

decision today may make that task more difficult.30 But the State also has a

compelling interest in prosecuting murders, rapes, armed robberies, and a whole

host of other serious crimes, and the right to be free from compelled self-

incrimination does not wax or wane based on the severity of a defendant’s

alleged crimes.

       This Court cannot change the Georgia Constitution, even if we believe

there may be good policy reasons for doing so; only the General Assembly and

the people of Georgia may do that. And this Court cannot rewrite statutes. This

decision may well have implications for the continuing validity of the implied

consent notice as applied to breath tests, but revising that notice is a power

reserved to the General Assembly. Having considered the text of Paragraph XVI

and the context in which it was enacted, as well as all of the arguments made by

the parties and the amici, we conclude that Paragraph XVI precludes admission



       30
         Although as we discussed in Olevik, this Paragraph XVI analysis is limited to breath
tests. See Olevik, 

, 233 n.2 (“Nothing we say here should be understood as
casting any doubt on Strong’s” holding that Paragraph XVI was not implicated by a blood
test.).



                                             90
of evidence that a suspect refused to consent to a breath test.31 Consequently, we

conclude that OCGA §§ 40-5-67.1 (b) and 40-6-392 (d) are unconstitutional to

the extent that they allow a defendant’s refusal to submit to a breath test to be

admitted into evidence at a criminal trial.

      We reverse the trial court’s denial of Elliott’s motion to suppress.

      Judgment reversed. All the Justices concur, except Warren and Ellington,

JJ., disqualified.




      31
          To the extent Elliott also raises arguments under OCGA § 24-5-506 and her
constitutional right to due process, we need not address them given our resolution of her
argument under Paragraph XVI.

                                           91
      BOGGS, Justice, concurring.

      I join the Court’s opinion in full, but write to clarify certain implications

of the Court’s decision today and our earlier decision in Olevik v. State, 

 (

) (2017).

      First, it is important to identify the provisions of the implied consent law

that are not affected. As acknowledged both in the Court’s opinion today and in

Olevik, the scope of these decisions is limited to chemical tests of a driver’s

breath; they do not apply to tests of a driver’s blood. Also unaffected is the core

component of the implied consent enforcement scheme: the administrative

license suspension provided by OCGA § 40-5-67.1 (c) and (d). Additionally, the

holding that a driver’s refusal to take a breath test may not be used in a criminal

proceeding does not forbid its use in an administrative proceeding concerning

suspension of a driver’s license, and that core function of the implied consent

law remains in force, notwithstanding the Court’s opinions today and in Olevik.

      That being said, these decisions affect significant portions of the implied

consent law. The statements in the implied consent notices set forth in OCGA

§ 40-5-67.1 (b) — that Georgia law “requires” the driver to submit to breath
testing; that the “refusal to submit . . . may be offered into evidence” against the

driver at trial; and that the driver’s license “may be suspended” (as opposed to

“shall”) if the driver submits and the results indicate an alcohol concentration

above a prohibited level — are likely to become problematic in future cases as

a result of Olevik or the Court’s decision today.32 And it is conceivable that

these decisions could affect the admission of previous DUI convictions pursuant

to OCGA § 24-4-417 (a) (1). Moreover, officers are required to read the

warnings in a “substantively accurate” form. Sauls v. State, 

, 167-

168 (

) (2013), overruled in part on other grounds, Olevik, 

 (2) (c) (iv) n. 11; see OCGA § 40-5-67.1 (b) (“Such notice shall be read

in its entirety but need not be read exactly so long as the substance of the notice

remains unchanged.”). But only the General Assembly can change what they are

required to read. Thus, the General Assembly may wish to revise the provisions

of the implied consent law, particularly the content of the implied consent

notice.

       It is also worth noting that the General Assembly and the people could

       32
         In Olevik, we rejected the appellant’s facial challenges to the implied consent notice
as well as his particular as-applied challenge. But in doing so, we noted that he had pointed
out “deficiencies in the implied consent notice,” and further observed, “The General
Assembly may wish to amend the implied consent notice statute; if it does, among the
changes it may consider would be a clearer explication of the right to refuse testing, and a
more accurate articulation of the likelihood of license suspension.” 302 Ga. at 250 and n.14.

                                              2
reconsider Paragraph XVI itself. The Court’s opinions here and in Olevik

discuss at length Paragraph XVI’s application to incriminating acts of a

defendant. This understanding of Paragraph XVI has become something of an

outlier in state and federal constitutional interpretations of self-incrimination,

but, as this Court observed in Olevik and demonstrates more fully today, it has

a long history stretching back over more than a century of case law and through

several Georgia Constitutions. And, of course, the amendment of a

constitutional provision, particularly one of such long standing that has

generated a substantial body of case law, should not be undertaken lightly.

Moreover, the potential effects of such an amendment upon other areas of law

should also be borne in mind. However, because the Court’s decision today is

rooted in a constitutional provision that was adopted by the legislature and the

people, if the General Assembly and the people of Georgia see fit to take our

self-incrimination law in a different direction, a clear understanding of the scope

and impact of our decision here today may aid in informing their decision.

      I am authorized to state that Justice Blackwell and Justice Bethel join in

this concurrence.




                                        3
                          Decided February 18, 2019.

      DUI breath test; refusal; constitutional question. Clarke State Court.

Before Judge Auslander.

      Willis Law Firm, Gregory A. Willis, Jared T. Hall, for appellant.

      Carroll R. Chisholm, Jr., Solicitor-General,William W. Fleenor, Ethan M.

Makin, Assistant Solicitors-General; Kenneth W. Mauldin, District Attorney,

for appellee.

      Meg E. Heap, District Attorney; Miller & Key, J. Scott Key; William A.

Finch, Solicitor-General, David Holmes, Samuel R. d’Entremont, Assistant

Solicitors-General; Gilbert A. Crosby, Peter J. Skandalakis, Robert W. Smith,

Jr.; Luftman, Heck & Associates, Daniel J. Sabol; Huey Defense Firm, Donald

T. Huey, Blaise J. Katter; Christopher M. Carr, Attorney General, Patricia B.

Attaway Burton, Deputy Attorney General, Paula K. Smith, Senior Assistant

Attorney General, David S. McLaughlin, Senior Assistant Attorney General,

Jameson B. Bilsborrow, Assistant Attorney General, Andrew A. Pinson,

Solicitor-General, Ross W. Bergethon, Deputy Solicitor-General, amici curiae.




                                      4