MEDWED.31-6

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED:
THE PROSECUTORIAL CHARGING DECISION AND
THE INNOCENCE REVOLUTION
Daniel S. Medwed*

INTRODUCTION
Efforts to rectify wrongful convictions in the United States
arguably represent a new civil rights movement for the twenty-first
century.1 Since 1989, post-conviction DNA testing has exonerated over
two hundred and fifty inmates, their innocence proven beyond a shadow
of a doubt through science, and at least three hundred other innocent
prisoners have gained their freedom in cases lacking the magic bullet of
DNA.2 Studies of these cases reveal that specific factors tend to cause
wrongful convictions in the first place. Misbehavior by prosecutors—
especially involving the suppression of exculpatory evidence—has
emerged as one of those factors.3
This Symposium directly (and commendably) tackles the problem
of under-disclosure of evidence by prosecutors. Encouraging
prosecutors to adhere more closely to existing disclosure rules advances
the ends of fairness by boosting the capacity of the defense to prepare
* Professor of Law, S.J. Quinney College of Law, University of Utah. J.D. Harvard Law
School, 1995; B.A. Yale College, 1991. In sum and substance, this Article will appear as a
chapter in my forthcoming book on the topic of prosecutors and wrongful convictions to be
published by New York University Press. I would like to thank Ellen Yaroshefsky for inviting
me to participate in this Symposium and Chayce Clark for his helpful research assistance.
1 See, e.g., Press Release, Innocence Project, As 100th Innocent Prisoner Is Freed by DNA
Tests, Innocence Network Convenes to Map the Future of “New Civil Rights Movement”
in Criminal Justice, (Jan. 17, 2002), available at http://www.deathpenaltyinfo.org/
article.php?did=280&scid=1.
2 For a current listing of DNA exonerations in the United States, see The Innocence Project,
http://www.innocenceproject.org (last visited June 25, 2010). See also Samuel R. Gross et al.,
Exonerations in the United States, 1989-2003, 95 J. CRIM. L. & CRIMINOLOGY 523, 524, 533-35
(2005) (citing 196 non-DNA exonerations during this period, not including roughly 135 innocent
defendants framed and later freed as part of the Rampart scandal in Los Angeles and in Tulia,
Texas).
3 See, e.g., Ellen Yaroshefsky, Wrongful Convictions: It Is Time to Take Prosecution
Discipline Seriously, 8 D.C. L. REV. 275, 276, 278, 285 (2004).

2187

Electronic copy available at: http://ssrn.com/abstract=1545937

MEDWED.31-6

2188

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

for trial. Increased disclosure can also bolster the accuracy of criminal
adjudications by minimizing the risk that innocent criminal defendants
will be wrongfully convicted; armed with exculpatory and other
evidence, the innocent are better positioned to reject plea offers and
mount solid defenses at trial. In short, more disclosure in more cases in
a more timely fashion is a good idea.
Yet any discussion about prosecutorial disclosure is incomplete
without paying some attention to an equally vital moment in the pretrial
process: the initial decision to charge a suspect with a crime
whatsoever. The mere decision to charge tends to set in motion a
sequence of events that inexorably result in either a plea offer or a
trial—even where the case is weak and where the prosecution has
complied fully with its disclosure duties.4 This Article grapples with the
topic of prosecutorial charging decisions in light of the “Innocence
Revolution.”5 Part I of the Article explores the rules and practices
surrounding prosecutorial charging decisions, pointing out some of the
flaws in this regime that may accidentally lead to charging innocent
suspects with crimes. Part II proposes a series of modest reforms to the
charging process designed to reduce the possibility that the innocent
will face criminal charges at all.
I. THE CHARGING DECISION
Although the police typically initiate the criminal process by
investigating alleged criminal activity and making arrests, prosecutors
enjoy the largely unbridled discretion to determine what, if any, charges
are actually filed against a defendant. In general, prosecutors may not
file a criminal charge unless it is supported by “probable cause” to
believe in the person’s guilt.6 This standard is inherently quite minimal;
it only requires enough evidence for the individual prosecutor to
subjectively think it is more likely than not that the person committed
the crime.7 Further diluting the intrinsic weakness of this standard are
4 See Andrew D. Leipold, How the Pretrial Process Contributes to Wrongful Convictions, 42
AM. CRIM. L. REV. 1123, 1160 (2005) (“[O]nce the process against an innocent suspect begins,
there is little chance that a case will be derailed because of a lack of evidence.”).
5 See Mark A. Godsey & Thomas Pulley, The Innocence Revolution and Our “Evolving
Standards of Decency” in Death Penalty Jurisprudence, 29 U. DAYTON L. REV. 265, 267 (2004)
(“The lessons of the Innocence Revolution begin with the realization that our system is not as
accurate as we believed even 10 years ago.”); Lawrence C. Marshall, The Innocence Revolution
and the Death Penalty, 1 OHIO ST. J. CRIM. L. 573, 573 (2004) (“Spawned by the advent of
forensic DNA testing and hundreds of post-conviction exonerations, the innocence revolution is
changing assumptions about some central issues of criminal law and procedure.”).
6 See, e.g., MODEL RULES OF PROF’L CONDUCT R. 3.8(a) (2006).
7 In fact, it appears as if a mere fifty percent chance of guilt may suffice to justify a charging

Electronic copy available at: http://ssrn.com/abstract=1545937

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2189

various doctrinal rules suggesting that (1) only the government’s
evidence is included in this calculus without reference to the defense’s
claims, (2) the credibility (or lack thereof) of the government’s
witnesses is not worthy of consideration, and (3) legally inadmissible
hearsay may be taken into account.8
Granted, the fact that a person may be charged with a crime due to
the presence of probable cause does not mean that a prosecutor must do
so. Prosecutors possess ample discretion in this area of their work, and
for good reason. The need for individualized justice, the problem of
“overcriminalization” in our criminal codes, and the finite resources of
law enforcement agencies all augur prosecutorial discretion in crafting
charges.9 Not everyone should be prosecuted for every crime they
apparently committed. Such a practice would lead to an overabundance
of prosecution, drain government assets, and impose penalties for
outmoded or ill-defined crimes that regrettably remain on the books.10
Still, there must be some check on prosecutorial discretion as a matter
of due process and fundamental justice, particularly to protect the
actually innocent. Model ethical codes, as well as state-specific rules,
offer prosecutors a modicum of guidance in deciding whether to file
charges in cases where the barebones probable cause standard has been
met.11 Prosecutors are generally urged to consider the defendant’s role
in the crime, his background and criminal history, his willingness to aid
the government in developing a case against another transgressor, and
the impact on the victim in the charging determination.12 Prosecutors
should also evaluate the availability of noncriminal dispositions in their
decisions and need not present all charges supported by the evidence, let
alone the highest ones.13 Several ethics codes, moreover, forbid
prosecutors from “overcharging” solely in the hopes of developing

decision. See Fred C. Zacharias & Bruce A. Green, The Duty to Avoid Wrongful Convictions: A
Thought Experiment in the Regulation of Prosecutors, 89 B.U. L. REV. 1, 9 (2009).
8 See, e.g., R. MICHAEL CASSIDY, PROSECUTORIAL ETHICS 15 (2005).
9 See BENNETT L. GERSHMAN, PROSECUTORIAL MISCONDUCT 155-56 (2007); Leslie C.
Griffin, The Prudent Prosecutor, 14 GEO. J. LEGAL ETHICS 259, 263-66 (2001).
10 See GERSHMAN, supra note 9; see also Griffin, supra note 9, at 263-64 (“[S]ome analysts
have argued that prosecutors are more suited than the legislature to adapt the criminal law to new
circumstances and to identify when the prosecution of certain statutes would be anachronistic.”).
11 See Bennett L. Gershman, A Moral Standard for the Prosecutor’s Exercise of the Charging
Discretion, 20 FORDHAM URB. L.J. 513, 513 (1993) (“[Although] legal, political, experiential, and
ethical considerations inform and guide the charging decision. . . . no subject in criminal law is as
elusive as that of prosecutorial discretion in the charging process.”).
12 See, e,g., Amie N. Ely, Note, Prosecutorial Discretion as Ethical Necessity: The Ashcroft
Memorandum’s Curtailment of the Prosecutor’s Duty to “Seek Justice,” 90 CORNELL L. REV.
237, 242-48 (2004); see also ANGELA J. DAVIS, ARBITRARY JUSTICE: THE POWER OF THE
AMERICAN PROSECUTOR 23 (discussing some of the reasons why prosecutors may forgo filing
charges).
13 See Ely, supra note 12, at 245-48.

MEDWED.31-6

2190

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

leverage for plea bargaining negotiations.14
Most notably, a
prosecutor’s individual doubts about the guilt of the accused should be
taken into account.15 Yet for all the lofty rhetoric espoused in the
canons of ethics, it remains unclear how this discretion is exercised in
reality.
Charging decisions, like much of the grist of the prosecutorial mill,
occur behind the scenes. This makes it difficult to gauge the degree to
which prosecutors rely on suspect choices as part of their initial
formulation of criminal charges.16 Prosecutors’ offices handle charging
decisions in many different ways. In most jurisdictions, the chief
prosecutor is elected by the public and then proceeds to appoint a
number of assistant attorneys to oversee the day-to-day operations of
the office. Some chief prosecutors vest almost complete charging
discretion in the hands of their assistants; others establish specific
charging guidelines for certain crimes and may even require assistants
to seek permission from a superior before deviating from them in a
particular case.17 On the whole, however, chief prosecutors tend to give
significant autonomy to line assistants in rendering charging decisions
with little supervision and even less accountability.18
What is more, charging decisions receive limited scrutiny from
those outside prosecutorial enclaves because the judicial and legislative
branches normally defer to their executive branch counterparts—
prosecutors—when it comes to charging crimes. Prosecutorial charging
decisions are essentially exempt from judicial review on two grounds:
(1) because courts lack the expertise and access to evidence to secondguess these choices,19 and (2) due to separation of powers concerns.
Simply put, judges appear hesitant to question executive department law
14 See CASSIDY, supra note 8, at 18-19. Scholars perceive the practice of “overcharging” to
be rampant in the criminal justice system. See, e.g., DAVIS, supra note 12, at 31 (“Prosecutors
routinely engage in overcharging, a practice that involves ‘tacking on’ additional charges that
they know they cannot prove beyond a reasonable doubt or that they can technically prove but are
inconsistent with legislative intent or otherwise inappropriate.”).
15 See, e.g., STANDARDS FOR CRIMINAL JUSTICE: PROSECUTION FUNCTION & DEF.
FUNCTION § 3-3.9(b)(i) (3d ed. 1993).
16 See, e.g., Patrick J. Fitzgerald, Thoughts on the Ethical Culture of a Prosecutor’s Office, 84
WASH. L. REV. 11, 12 (2009) (“The fact is that more of a prosecutor’s important work takes place
behind closed doors than in public.”).
17 See DAVIS, supra note 12, at 33-34; see also Kenneth J. Melilli, Prosecutorial Discretion
in an Adversary System, 1992 BYU L. REV. 669, 683 (1992).
18 See DAVIS, supra note 12, at 33-34; see also Melilli, supra note 17, at 687 (discussing how
offices tend to screen cases and how, even where supervisors make an initial screening
assessment, that assessment is often made quite quickly and the line assistant subsequently
assigned to a case often defers to the initial evaluation).
19 See CASSIDY, supra note 8, at 23-24. To be sure, prosecutors may not arbitrarily select a
defendant for prosecution without running afoul of equal protection principles. See GERSHMAN,
supra note 9, at 164-83. Similarly, defendants enjoy constitutional protections from prosecutors
possessing a “vindictive” personal motive for pursuing charges against them. Id. at 183-206.

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2191

enforcement decisions before they reach fruition in court.20 The
legislative branch also provides meager oversight of prosecutorial
charging decisions. Although in theory, legislatures participate in the
charging endeavor by defining and amending criminal offenses in a
particular jurisdiction, in practice, legislatures have long since abdicated
responsibility for charging to prosecutors.21 As a result, not only do
prosecutors have vast discretion in determining whether to charge a
person with a crime at the outset, but that discretion is subject to token
judicial review and scant accountability to the legislature.
Compounding the lack of transparency surrounding initial charging
decisions by prosecutors is the nature of the process through which
many criminal charges ultimately make it to court. The procedure for
filing formal criminal charges varies considerably from jurisdiction to
jurisdiction, but shares some common features. After a person is
arrested by the police, he must appear before a judge or magistrate
normally within forty-eight hours to determine whether he will be
released or detained pending a trial date.22 Prosecutors in some
jurisdictions file a charging document at this hearing, often referred to
as an “information,” that outlines the nature of the accusations based on
the contents of the arrest file.23
In the federal system and in many states, though, criminal charges
(especially felonies) must first be presented to a grand jury for
indictment before proceeding to trial.24 It is often said that the grand
jury serves as a rubber stamp to validate prosecutorial charging choices
rather than as a bulwark against injustice.25 Indeed, the rules governing
grand jury practice undeniably favor the prosecution. First and
foremost, prosecutors dictate the flow of information to the grand jury.
The bulk of a grand jury’s efforts occur away from judges, defense

20

See United States v. Armstrong, 517 U.S. 456, 465 (1996); GERSHMAN, supra note 9, at

156.
21 See Robert L. Misner, Recasting Prosecutorial Discretion, 86 J. CRIM. L. & CRIMINOLOGY
717, 743 (1996) (“In the area of charging, prosecutorial decisions—such as whether to prosecute,
how long to sentence, and whether to dismiss charges—all contribute to the creation of the
prosecutor as the real policy-maker within the criminal justice system. At best the legislature is a
lesser partner whose role is to set the outer parameters of criminal law policy and to fund
prisons.”).
22 See DAVIS, supra note 12, at 24.
23 Id. The U.S. Supreme Court has not required that grand juries be convened at the state
level and, accordingly, over half of the states have selected other methods for charging cases,
such as the filing of informations and/or holding of probable cause hearings before judges. See
CASSIDY, supra note 8, at 26.
24 The United States Constitution provides that a criminal defendant in federal court has a
right to an indictment for all “capital, or otherwise infamous crime[s].” U.S. CONST. amend. V.
25 See, e.g., Gershman, supra note 11, at 520 (“Anybody familiar with the criminal justice
system knows that the grand jury does not act on its own and that the prosecutor controls grand
jury action.”).

MEDWED.31-6

2192

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

lawyers or the media. In fact, defense counsel is typically barred from
the room.26 Second, prosecutors may present inadmissible evidence in
making their case,27 and they generally are not required to present
evidence that exculpates the defendant.28 Third, grand juries usually
may issue an indictment when a bare majority of its members finds
probable cause to believe in the defendant’s guilt.29 Overall, a strong
presumption of deference, even correctness, for prosecutorial charging
decisions animates the early stages of the criminal process and shrouds
those choices in a veil of secrecy that is rarely lifted for public view.30
Nevertheless, dubious charging decisions involving innocent
suspects do not exist purely in the world of hypothetical cases. Several
high-profile prosecutorial blunders in recent years, including the
notorious Duke lacrosse case, illustrate how criminal charges ostensibly
based on probable cause were later revealed to be erroneous. Had the
flaws in the Duke case not been revealed promptly, a handful of
promising college students might have been wrongfully convicted.
The basic facts of the Duke lacrosse incident are well-known, yet
still warrant a brief discussion here in order to highlight the risks posed
by the rules associated with prosecutorial charging decisions. On
March 13, 2006, three Duke University students hosted a party in
Durham, North Carolina.31 Most of the roughly forty students in
attendance were members of the men’s lacrosse team. At the request of
one of the party’s hosts, a local escort service dispatched two AfricanAmerican exotic dancers, Crystal Gail Mangum and Kim Pittman, to the
location. Pittman arrived by herself and waited quite a while before
Mangum, whom she did not know well, appeared. Witnesses described
Mangum as being unsteady on her feet throughout her time at the
party.32
Shortly before midnight, the two dancers began their performance
in exchange for $400 each. Mangum, still wobbly, fell to the ground.
The performance evoked sexual remarks from the attendees, with one
partygoer raising a broom and recommending it to the dancers for their
sexual pleasure. Upset by this comment, Mangum and Pittman abruptly
stopped dancing, left the room, and went toward the back of the house.
They were followed by several students seeking to apologize for the
26
27
28
29
30
31

See CASSIDY, supra note 8, at 26-29.
See id.
See DAVIS, supra note 12, at 24.
See CASSIDY, supra note 8, at 26-29.
See, e.g., United States v. Armstrong, 517 U.S. 456, 464 (1996).
The description of the incident that gave rise to the Duke lacrosse rape case stems from
Robert P. Mosteller, The Duke Lacrosse Case, Innocence, and False Identifications: A
Fundamental Failure to “Do Justice,” 76 FORDHAM L. REV. 1337, 1342-46 (2007).
32 Id.

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2193

broom incident and requesting a longer performance, one, in their view,
more commensurate with the fee. The dancers went to the bathroom
where they had left their belongings and remained there for a period of
time. Some guests left at this point; others continued to decry having
paid money for a brief performance. Still others were simply hanging
around the house.33
The dancers eventually left the house and went to their car, only to
return a short time later and sequester themselves in the bathroom.
Within minutes, Mangum and Pittman left the bathroom and the house
yet again. But, rather than returning to the car, Mangum lingered
outside and seemingly engaged in an odd pattern of behavior—banging
on the house door and requesting to be re-admitted, struggling to
maintain her balance, attempting unsuccessfully to have coherent
conversations with assorted guests, and lying on her back on the porch.
One of the guests ultimately assisted Mangum to Pittman’s car and, as
Pittman prepared to drive off with Mangum in tow, a series of racial
epithets were exchanged between Pittman and the attendees. Pittman
drove off and placed a 911 call to report that a number of white men
were inflicting racial barbs at bystanders.34
Pittman drove to a grocery store, but Mangum refused to exit the
car and appeared to be unconscious. That prompted Pittman to place
another 911 call and, at the direction of the police, to take Mangum to
Durham Center Access where she was seen by a nurse. The nurse asked
Mangum, as a matter of standard operating procedure, whether she had
been raped; in response, Mangum indicated—for the first time—that
she had been victimized in that fashion. She later recanted that
assertion at the Duke Medical Center’s emergency room before
changing her mind and reviving her claim that she had in fact been
raped. The police decided to treat it as a potential rape case.35
Durham County District Attorney Mike Nifong took charge of the
case almost immediately after he learned about it in late March 2006,
and began to make public statements to the media that cast the
prospective defendants in a poor light. Indeed, due in part to Nifong’s
willingness to bask in the media spotlight, the investigation soon
became a national cause célèbre fraught with racial and socioeconomic
overtones. The media at first largely depicted the case as a “towngown” affair—a sordid tale of spoiled, white student-athletes from an
elite university taking advantage of an impoverished and troubled black
woman. Nifong, in essence, positioned himself as the knight in shining
armor riding roughshod over Duke royalty in order to right a grievous
33
34
35

Id.
Id. at 1344-45.
Id. at 1345-46.

MEDWED.31-6

2194

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

injustice.36
In early April 2006, Mangum identified three Duke lacrosse
players—Dave Evans, Colin Finnerty, and Reade Seligmann—as her
assailants. Some of the physical evidence corroborated Mangum’s
claim that she had suffered a sexual assault. The doctor who initially
examined Mangum in the early morning of March 14 had detected
vaginal swelling and a nurse characterized Mangum’s behavior at the
time as consistent with that of a victim of a sex crime. Finally, Evans’s
DNA was partially matched to DNA found on a fake fingernail that
belonged to Mangum and that was recovered from the bathroom where
she claimed the rape occurred. 37
Based on that evidence alone, could Nifong have legitimately
found probable cause to believe that a rape had happened and that
Evans, Finnerty, and Seligmann were the culprits? It is quite possible.
But, the addition of other evidence to the mix—most of which Nifong
knew about in early April—complicates the matter. Here are the critical
“bad” facts in the prosecution’s case, in no particular order:
• Mangum had given a number of conflicting statements
about the event;
• Her medical records revealed a history of severe mental
health problems, including bipolar disorder;
• No DNA evidence from the rape kit matched the three
alleged perpetrators;
• DNA from men other than any of the players was found in
the rape kit;
• Pittman had at one occasion stated that Mangum’s story
was “a crock”; and
• In the first identification procedure, Mangum had identified
Seligmann only as a guest at the party, not as a participant
in the assault.38
Despite the conflicting evidence, Nifong proceeded to file rape
charges against the three men and successfully returned indictments
against all of them. Before trial, however, the North Carolina State Bar
filed a disciplinary complaint against Nifong related to his pretrial
statements. Nifong asked the North Carolina Attorney General’s Office
to take over the case while the disciplinary complaint was pending.
Special prosecutors from that office began to investigate the case anew;
by December, they concluded that Evans, Finnerty and Seligmann were
innocent and ordered the dismissal of the charges. The state bar found
Nifong guilty of serious ethical lapses stemming from his improper
36
37
38

Nifong was sanctioned for his contribution to improper pretrial publicity. Id. at 1348-50.
For a discussion of the evidence against the three defendants, see id. at 1373.
Id. at 1372-73.

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2195

statements to the media and his failure to disclose exculpatory DNA
evidence to the defense, as well as other mistakes regarding the
discovery process. He was ultimately disbarred.39
One grave misstep by Nifong, though, never resulted in an ethical
complaint: his decision to charge the three players with rape and seek an
indictment in the first place. North Carolina, like many other states,
provides only that a prosecutor shall “refrain from prosecuting a charge
that the prosecutor knows is not supported by probable cause.”40
Professor Robert Mosteller, who has studied the Duke lacrosse case
extensively, considers this standard “very limited,” which may explain
why the state bar never made allegations against Nifong under this
provision.41 In his testimony before the Disciplinary Hearing
Commission, Nifong intimated that he did not know the case lacked
probable cause to pursue it. He observed that the victim’s statement
that a rape occurred and that the three student-athletes committed it was
sufficient to get the case to a jury, irrespective of any inconsistent
evidence.42 If Nifong had been accused of improper charging by the
state bar, he would likely have escaped punishment given the deferential
ethical rule in this area.43
II. POTENTIAL REFORMS TO THE RULE REGIME GOVERNING
PROSECUTORIAL CHARGING DECISIONS
Mike Nifong’s outrageous choice to file criminal charges in the
Duke lacrosse rape case and the state bar’s reluctance, perhaps even
inability, to find ethical violations for this behavior show the failings of
the current rule system governing prosecutorial charging decisions. On
the one hand, prosecutors deserve some freedom to strike individuallytailored charging decisions and to channel their energies to the cases
they deem most meritorious. On the other hand, this freedom should
not be wholly unfettered; the duty to “do justice” encompasses the
obligation not to prosecute the factually innocent.44 To mitigate the risk
39 Id. at 1348-64; see also Zacharias & Green, supra note 7, at 12 (praising disbarment and
“meaningful sanctions” in the Nifong case as “an exceptional instance”).
40 REVISED RULES OF PROF’L CONDUCT OF THE N.C. STATE BAR R. 3.8(a) (2007).
41 Mosteller, supra note 31, at 1368.
42 Id. at 1373-74.
43 Id. at 1372; see also id. at 1376 (“That Nifong knew no probable cause existed in this case
is not clearly established. However, that he had abundant notice of the problematic nature of the
case is undeniable.”).
44 Id. at 1366 (“However that duty [to do justice] is defined and whatever its precise origins,
all agree that the prosecutor has a special duty not to prosecute the innocent.”); see also Melilli,
supra note 17, at 672 (“[M]any prosecutors do, and all should, regard the possibility of charging
an innocent person as ‘the single most frightening aspect of the prosecutor’s job.’”).

MEDWED.31-6

2196

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

that innocent defendants will suffer the taint of an unjust charge and
face the prospect of a subsequent wrongful conviction,45 there must be
greater accountability for prosecutors in making initial charging
decisions and greater transparency in the decision-making process itself.
What reforms might result in improved accountability and transparency
in the charging process? Upon reflection, two possibilities come to the
fore: (1) altering the evidentiary threshold that prosecutors must
overcome to file charges at the outset, and (2) spurring prosecutors to
develop internal review committees equipped with “devil’s advocates”
to pre-screen factually weak cases.
A.

Altering the Charging Standard

As discussed above, prosecutors in the vast majority of
jurisdictions may file criminal charges so long as they believe they are
supported by probable cause, a standard that many scholars have
derided as woefully inadequate in filtering out the innocent.46 One
possible change to this rule entails increasing the quantum of proof from
probable cause to something more, such as sufficient evidence to obtain
a conviction.47 Professor Bennett Gershman contends that “the
45 Even where a charge does not produce a conviction, there are enormous costs inflicted
upon a defendant. See Angela J. Davis, The American Prosecutor: Independence, Power, and the
Threat of Tyranny, 86 IOWA L. REV. 393, 413 (2001) (“[T]he defendant might labor under the
shadow of suspicion that often lingers even after charges are dismissed or unproven.”); Melilli,
supra note 17, at 672 (“[T]he mere filing of a criminal charge can have a devastating effect upon
an individual’s life, including potential pretrial incarceration, loss of employment, embarrassment
and loss of reputation, the financial cost of a criminal defense, and the emotional stress and
anxiety incident to awaiting a final disposition of the charges.”).
46 Kenneth Melilli, for example, has condemned the probable cause standard as “little more
than heightened suspicion and it is not even remotely sufficient to screen out individuals who are
factually not guilty.” Melilli, supra note 17, at 680-81; see also Bruce A. Green, Prosecutorial
Ethics as Usual, 2003 U. ILL. L. REV. 1573, 1584 n.53 (observing that the “‘probable cause’
standard is widely considered too low”). One example of an institution’s choosing to hold itself
to a higher charging standard is the United States Attorney’s Office, the chief federal
prosecutorial agency, whose office manual prescribes that “[t]he probable cause standard is . . . a
threshold consideration only. Merely because this requirement can be met in a given case does
not automatically warrant prosecution; further investigation may be warranted, and the prosecutor
should still take into account all relevant considerations . . . .” U.S. ATTORNEYS’ MANUAL
§ 9-27.200(B) (U.S. Dep’t of Justice 1997), available at http://www.justice.gov/usao/eousa/
foia_reading_room/usam/title9/27mcrm.htm#9-27.200; see also Stephanos Bibas, Prosecutorial
Regulation Versus Prosecutorial Accountability, 157 U. PA. L. REV. 959, 1003 (2009) (describing
how the New Jersey Attorney General drafted statewide guidelines to govern charging enhanced
drug penalties).
47 See Mosteller, supra note 31, at 1368; see also Mellili, supra note 18, at 701 (“[T]he
conscientious prosecutor, in zealous pursuit of society’s interest in justice, does not and should
not pursue cases unless personally satisfied beyond a reasonable doubt of the defendants’ guilt.”);
id. at 681-82 (discussing the standards proposed by the U.S. Department of Justice, National
District Attorneys Association and the American Bar Association, each of which essentially

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2197

prosecutor should engage in a moral struggle over charging decisions,
and should not mechanically initiate charges.”48 Specifically, Gershman
has argued that “responsible prosecutors should be morally certain that
the defendant is guilty,”49 or that prosecutors should only proceed with
charges if they are “personally convinced of the defendant’s guilt.”50
Although Gershman’s arguments are compelling, demanding nearly
conclusive personal belief of guilt might admittedly have a downside.
Among other effects, it could put too much pressure on prosecutors to
work extensively as fact-finders at the front end of every criminal case,
siphon off precious resources, and allow fewer cases to go to trial. Even
so, lifting the standard from probable cause to a level that comes closer
to approximating the evidentiary threshold for establishing guilt at trial
(proof of guilt beyond a reasonable doubt) would certainly help weed
out borderline cases and, with them, undoubtedly save some actually
innocent suspects.
Another reform that could accompany the installation of a higher
charging standard, or even prove useful under current charging norms,
is to require prosecutors to weigh any exculpatory evidence, such as the
defense’s explanation of the events, against the incriminating facts. It
strikes me as illogical that prosecutors are ethically obliged only to
credit the facts implicating the suspect in deciding whether a charge is
warranted. How can a prosecutor produce a genuine charging decision
while artificially excising the “bad” facts from the equation?
Shifting the perspective from which probable cause is determined
under the current charging standard could also reap rewards. At
present, the standard in almost every jurisdiction is entirely subjective.
As long as a prosecutor does not “know” that the charges are inadequate
(namely, that they lack probable cause), she is not obliged to “refrain”
from filing. This naturally creates an incentive for prosecutors to take
the evidence at face level and not delve any deeper into the intricate
minutiae of a case: to put the matter effectively into the jury’s hands.
Several commentators defend this standard on the basis that determining
guilt is a jury function and that victims deserve their day in court.51
recommends a standard of sufficient evidence to support a conviction).
48 Gershman, supra note 11, at 522; cf. Green, supra note 46, at 1589. Green proposes an
intermediate standard whereby “charges should not be brought unless the prosecutor reasonably
believes that the accused is guilty of the crimes charges.” Id.
49 Gershman, supra note 11, at 524.
50 Bennett L. Gershman, The Prosecutor’s Duty to Truth, 14 GEO. J. LEGAL ETHICS 309, 338
(2001).
51 See, e.g., H. Richard Uviller, The Virtuous Prosecutor in Quest of an Ethical Standard:
Guidance from the ABA, 71 MICH. L. REV. 1145, 1168 (1973) (insisting that, while prosecutors
should “protect the innocent,” they should never serve as the “sole arbiter of truth and justice”);
id. at 1159 (“Thus, when the issue stands in equipoise in his own mind, when he is honestly
unable to judge where the truth of the matter lies, I see no flaw in the conduct of the prosecutor

MEDWED.31-6

2198

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

Entrusting the resolution of guilt or innocence to the jury appeals to
populist impulses and, more crassly, it allows prosecutors to shed
ultimate responsibility for the outcome. The statement “the jury has
spoken” is a common catchphrase after a shocking or disappointing
result at trial. Although showing faith in the jury system resonates with
our conception of democracy and the desire to be judged by one’s peers,
I have misgivings about prosecutors “passing the buck” in weak cases
under the guise of populism. Juries are unreliable evaluators of guilt or
innocence, as evidenced by the hundreds of documented wrongful
convictions emanating from jury trials, and the deliberation room is just
as often a forum for prejudice and emotion as it is a sanctuary of
reasoned, dispassionate debate.52 The irresponsibility of using populist
arguments about jury deference to justify specious charging decisions is
further substantiated by the reality that most matters are resolved
through plea bargains and never see the inside of a trial court.53 As
repeat players in the field with access to vast sources of information,
prosecutors are well-positioned to serve as front-end gatekeepers
preventing weak cases from entering the litigation process at all.54
Accordingly, one might consider altering the focus of the probable
cause test from subjective to objective: asking whether a “reasonable
prosecutor” would have found probable cause to file charges. An
upshot of this change would be to encourage prosecutors to examine the
evidence more rigorously; lack of knowledge about the weakness of the
case would no longer comprise an acceptable excuse. Several states—
California, Hawaii, Illinois, Iowa, Maine, and New York—have added
an objective twist to the subjective standard for evaluating charging
decisions.55 Prosecutors in those states, with minor variations, should
“not institute or cause to be instituted criminal charges when [the
prosecutor] knows or reasonably should know that the charges are not
supported by probable cause.”56 Assuming that injecting an objective
component to the probable cause formulation is a reform well-worth
contemplating, the question then revolves around when a prosecutor
“reasonably should know” that a charge is deficient. Should a
reasonable prosecutor be required to engage in an independent
who fairly lays the matter before the judge or jury.”).
52 See, e.g., Melilli, supra note 17, at 700. Acquittal rates in the early twenty-first century
have reached unprecedented lows. Russell D. Covey, Fixed Justice: Reforming Plea Bargaining
with Plea-Based Ceilings, 82 TUL. L. REV. 1237, 1238-39 (2008).
53 See, e.g., Melilli, supra note 17, at 700.
54 Bennett Gershman has argued that prosecutors far too often pass responsibility for close
questions on to the jury and that, instead, each prosecutor should more actively serve as “a
gatekeeper of justice” given the risk of jury error. See Gershman, supra note 11, at 521.
55 See Mosteller, supra note 31, at 1367 n.140.
56 Some of those jurisdictions use the phrase “or it is obvious” instead of “or reasonably
should know.” Id.

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2199

investigation and analysis of the case in order to clear this hurdle? If so,
would those efforts be redundant of those of police?
A “reasonable prosecutor” could conceivably be required to
investigate the facts of each case prior to charging,57 but crafting such a
broad, wholesale obligation is impractical. Prosecutorial resources are
limited; inflicting a pre-charging duty on prosecutors to investigate
cases might serve mainly to burden the already-taxed coffers of these
agencies and duplicate the efforts of the police.58 Recognizing the
practical flaws with a blanket prosecutorial duty to independently
investigate cases in the pre-charge phase, Mosteller recommends
confining this duty to “high-profile” or “problematic” cases, for
instance, those akin to the Duke lacrosse case.59 Mosteller’s plan is
laudable, yet it could face practical difficulties, specifically in terms of
how to ascertain whether a particular case fits into this rarefied
category. More worrisome is the notion that defendants in certain highprofile cases—presumably those involving wealthy or notorious parties
or otherwise possessing salacious details—would receive greater
protection than others. Regardless of the merits of imposing a
prosecutorial duty to investigate in some form prior to charging, merely
adding an objective part to the charging test would help. At a
minimum, it would signal that prosecutors must critically examine the
evidence, if not investigate afresh, in cases where an initial review has
not given them an objectively reasonable sense of the case’s strength.
Charging decisions by prosecutorial ostriches who hide their head in the
sand and refuse to acquire the “knowledge” that a case is wanting would
no longer pass ethical muster.
For these proposed changes to the ethical rules to be effective,
however, state disciplinary bodies must be willing to enforce them.60
Sanctioning prosecutors more frequently for charging errors would
likely give many prosecutors greater pause before proceeding. Greater
pause, in turn, probably means a greater number of prosecutors
declining to charge. Recall the Duke lacrosse case. Noticeably absent
from the litany of ethical violations alleged by the North Carolina State
Bar against Mike Nifong was one relating to the fundamental choice to
charge the three students with rape. If there had been a realistic chance
57 Several scholars advocate creating some semblance of an investigative obligation for
prosecutors prior to rendering a charging decision. See id. at 1369-70 nn.148-150.
58 Id. at 1370-71.
59 Id. at 1370-79.
60 This is no easy feat. Disciplinary agencies have long displayed reluctance to chastise
prosecutors for misdeeds, setting their sights instead on private lawyers. See Yaroshefsky, supra
note 3, at 278-80 (discussing studies showing that prosecutorial misconduct is seldom subject to
disciplinary sanction); Fred C. Zacharias, The Professional Discipline of Prosecutors, 79 N.C. L.
REV. 721, 722 (2001) (“Numerous commentators have reacted by noting the dearth of cases in
which disciplinary authorities have sanctioned prosecutors.”).

MEDWED.31-6

2200

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

that the North Carolina State Bar would impose sanctions for misguided
charging decisions, would Nifong have wavered before making this
crucial decision? In other cases where flaws with the prosecution’s
theory were not revealed in equally timely fashion, would innocent
suspects have been spared the agony of going to trial and facing
conviction?61 Inspiring disciplinary boards to investigate and discipline
prosecutors for imprudent charging decisions would serve a key
deterrent function and nicely complement any alterations to the rules
surrounding the charging standard itself.
All told, (1) raising the evidentiary threshold for filing criminal
charges from probable cause of guilt to something closer to, yet less
than, proof beyond a reasonable doubt, (2) including defense evidence
in the calculus, and (3) adding an objective prong to the test to deter
prosecutors from acting like proverbial ostriches inclined to ignore the
holes in their cases would minimize the chance that innocent defendants
will be subject to criminal charges and ultimately convicted. That being
said, there should be further checks at the charging stage to safeguard
the innocent, one of which I will now address—the need for prosecutors
to establish internal “pre-screening” committees to review inherently
weak cases.
B.

Screening Committees to Review Suspect Cases

To decrease the chance that fundamentally questionable criminal
charges might technically survive even under a modified rule regime, I
would propose creating a secondary review structure within each
prosecutorial office: internal review committees to evaluate charging
decisions in cases that contain certain hallmarks of innocence. Internal
review committees are necessary because prosecutors, as wellintentioned as they may be, suffer from innately human cognitive biases
that deter them from rationally reviewing the evidence against a
potential suspect with the requisite equanimity and distance.62
Prosecutors usually enter a case after the police have investigated the
61 Prosecutorial misconduct surfaces frequently in studies as a factor in wrongful convictions.
See Yaroshefsky, supra note 3, at 278 (mentioning one study by the Innocence Project that cited
prosecutorial misconduct as a feature in forty-five percent of the first seventy-four DNA
exonerations).
62 Several scholars have suggested that prosecutors, like all humans, are not purely rational
actors but, rather, suffer from “bounded rationality.” See Alafair Burke, Improving Prosecutorial
Decision-Making: Some Lessons of Cognitive Science, 47 WM. & MARY L. REV. 1587, 1590-91
(2006). As Burke suggests, “[t]he potential for cognitive bias to creep into prosecutorial
decision-making starts from the earliest case-screening stages, when prosecutors must determine
whether sufficient evidence exists to proceed with a prosecution.” Id. at 1603.

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2201

crime and arrested the chief suspect. The phenomenon of police “tunnel
vision” has often already reared its ugly head by that point. Police
tunnel vision occurs when detectives, after concentrating on a particular
suspect, overestimate the evidence against that person and
subconsciously disregard the possibility of alternative perpetrators or
exculpatory evidence throughout the remainder of the investigation.63
What might explain this tendency? Psychologists attribute tunnel vision
mainly to a series of intrinsic cognitive biases that affect how people
perceive information and how they interpret what they have perceived.
In particular, Professors Keith Findley and Michael Scott suggest that
the foundation of tunnel vision in the criminal justice system lies in an
“expectancy” or “confirmation” bias.64 They observe that “[w]hen
people are led by circumstances to expect some fact or condition (as
people commonly are), they tend to perceive that fact or condition in
informationally ambiguous situations. This can lead to error biased in
the direction of the expectation.”65 That is, after a person initially
develops a theory about a topic, this bias may spur that individual to
selectively process newfound information in a manner that confirms,
rather than challenges, the original hypothesis.66 People often not only
interpret information in a fashion that reinforces a pre-existing theory,
but also affirmatively seek to collect data that validates their hypothesis
and avoid accumulating evidence that undercuts it.67 Even when
confronted with information that thoroughly decimates their original
thesis, people may nonetheless rigidly cling to their initial viewpoint, a
tendency referred to as “belief perseverance.”68 Other cognitive biases
also intersect with the expectancy or confirmation bias to yield tunnel
vision.69
63 See generally Keith A. Findley & Michael S. Scott, The Multiple Dimensions of Tunnel
Vision in Criminal Cases, 2006 WIS. L. REV. 291 (2006) (providing an in-depth, step-by-step
exploration of the ways in which tunnel vision infects every phase of a criminal proceeding). For
a discussion of how cognitive biases affect prosecutorial decision-making in the post-conviction
context, see Daniel S. Medwed, The Prosecutor as Minister of Justice: Preaching to the
Unconverted from the Post-Conviction Pulpit, 84 WASH. L. REV. 35, 51-53 (2009).
64 See Findley & Scott, supra note 63, at 308; see also Alafair S. Burke, Neutralizing
Cognitive Bias, 2 N.Y.U. J.L. & LIBERTY 512, 517 (2007) (“Prosecutorial tunnel vision can be
viewed as the culmination of confirmation bias and selective information processing, the
inclination to search out and recall information that tends to confirm one’s existing beliefs, and to
devalue disconfirming evidence.”).
65 See Findley & Scott, supra note 63, at 307; see also Burke, supra note 62, at 1594-96
(discussing the confirmation bias).
66 See Burke, supra note 62, at 1596-99 (describing the concept of selective information
processing).
67 See Findley & Scott, supra note 63, at 309.
68 Burke, supra note 62, at 1599-1601 (presenting the concept of belief perseverance); Burke,
supra note 64, at 518-19 (same); Findley & Scott, supra note 63, at 314 (same).
69 For instance, after a person has emerged as the chief suspect, “the hindsight bias would
suggest that, upon reflection, the suspect would appear to have been the inevitable and likely

MEDWED.31-6

2202

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

In the context of police investigations specifically, once the police
embrace a particular theory of the case, the detectives may view all
subsequent evidence through the lens of this expectation. Needless to
say, this can produce some potentially distorted images. The flaws in
the case become incidental rather than fundamental. In addition, the
police may be eager to pursue avenues of investigation that substantiate
their view of guilt and disinclined to explore investigative leads that
support innocence.70 Worse yet, police may discount (or even reject)
exculpatory evidence that undermines their theory. All of this occurs
largely on a subconscious level. It is not as if the police purposefully
decide to sink the chief suspect and to ignore contrary evidence. Thus,
one of the most nefarious aspects of tunnel vision is that those deepest
in its throes may be those least aware of its existence.
The internal cognitive biases that give rise to tunnel vision in
police detectives are aggravated by external pressures as well.
Mounting calls by victims, the media, government officials, and police
chiefs to make an arrest do not fall on deaf ears; detectives are mindful
of the need to solve crimes, thereby signaling that the community is
safe. This may affect the detectives’ tactical decisions in investigating a
case and processing information about the likely perpetrator.71 By
interacting closely with a crime victim, the police may also become
emotionally attached to that person and his or her version of events.72
The sheer volume of their workload may further provoke the police to
channel their energies to the person they initially target as a suspect for
fear of stalling the resolution of that particular case and the start of work
on other pressing matters.73 Last, but not least, the method of gauging
police detective performance contributes to tunnel vision. Police
investigators are usually evaluated based on their “clearance rate,”

suspect from the beginning.” Findley & Scott, supra note 63, at 318. Hindsight bias often comes
equipped with a “reiteration effect” in which a person’s “confidence in the truth of an assertion
naturally increases if the assertion is repeated.” Id. at 319. Findley and Scott also point to a
related bias—outcome bias—as a factor in tunnel vision. Outcome bias relates to a person’s
retrospective assessment about the quality of a decision; if the outcome is bad, the person might
consider their previous decision as a bad one even though, of course, the outcome could not have
been known at the time of the decision. Id. at 319-20. Burke, for her part, suggests that the
“avoidance of cognitive dissonance” also informs prosecutorial decision-making. Burke, supra
note 62, at 1601-02. That is, she posits that people “desire to find consistency between one’s
behavior and beliefs.” Id. at 1601. The development of any “inconsistency between one’s
external behavior and internal beliefs creates an uncomfortable cognitive dissonance. To mitigate
the dissonance, people will adjust their beliefs in a direction consistent with their behavior.” Id.
70 Burke, supra note 62, at 1604 (“Confirmation bias will reduce the likelihood that the
investigation will be directed in a manner that would yield evidence of innocence.”).
71 See Findley & Scott, supra note 63, at 323-26.
72 Id. at 324-25.
73 Id. at 325; see also Leipold, supra note 4, at 1127 (discussing how deficiencies in the
evidence gathering process during an investigation can lead to wrongful convictions).

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2203

namely, the percentage of reported crimes that are eventually treated as
solved or, in the vernacular, “closed.”74 For obvious reasons, this
standard encourages frequently overworked detectives to operate
quickly to find suspects and arrest them as efficiently as possible, thus
feeding into the problem of tunnel vision.
The police investigation is where tunnel vision begins and where it
can generate the most damage given that all later stages of the process
build upon the information obtained by the police.75 But tunnel vision
is by no means the sole province of the police within the criminal justice
system. Prosecutors work closely with the police and may fall prey to a
comparable form of tunnel vision at the pretrial stage, especially as they
typically receive only the evidence implicating the person whom the
police consider the culprit.76 As Professor Randolph Jonakait has
observed, the trial prosecutor “does not see evidence about all the
possible suspects, but only the incriminating evidence concerning the
defendant. Not surprisingly, the picture presented to the prosecutor
almost always shows a guilty defendant.”77 This picture inevitably
becomes the reference point to which prosecutors look in assessing a
case, defining their expectations and laying the groundwork for tunnel
vision to fester.78 Also, studies show that the cognitive biases leading to
tunnel vision are worsened where key information is hidden or absent in
creating the initial expectation, as occurs in the transfer of case files
from the police to prosecutors.79 The police have the benefit of
entertaining (in theory) evidence inconsistent with a suspect’s guilt that
surfaces during the investigation of a crime before labeling a specific
person as the culprit. Access to that information is a luxury rarely
afforded to prosecutors.
The very nature of the prosecutor-police relationship also produces
incentives for prosecutors to take the outcome of the police
investigation as a fait accompli and to put on intellectual blinders to the
possibility of other outcomes. Prosecutors need a good working
relationship with the police, one based on trust and a smattering of
mutual respect, in order to do their jobs effectively. Prosecutors rely on
police to investigate cases, arrest perpetrators, and track down
74
75
76
77

See Findley & Scott, supra note 63, at 325-26.
Id. at 295.
Id. at 327-31.
Randolph N. Jonakait, The Ethical Prosecutor’s Misconduct, 23 CRIM. L. BULL. 550, 553
(1987).
78 See supra notes 63-68 and accompanying text. Another problem with the “picture”
contained in the police reports is that these reports often rely on “boilerplate” language that may
lack the specificity required to assess the individual merits of the case. See Erwin Chemerinsky,
The Role of Prosecutors in Dealing with Police Abuse: The Lessons of Los Angeles, 8 VA. J. SOC.
POL’Y & L. 305, 316 (2001).
79 See Findley & Scott, supra note 63, at 329-30.

MEDWED.31-6

2204

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

witnesses; police depend on prosecutors to validate those arrests by
securing convictions. By collaborating repeatedly with one another
over long periods, police and prosecutors may develop a shared
orientation toward “getting the bad guys” that creates a potent bond.80
Questioning the accuracy of a police investigation may jeopardize this
symbiotic (and co-dependent) relationship, not to mention imperil a
prosecutor’s ability to perform in future cases.81 Research shows that
failing to show group loyalty in general imposes profound costs on the
perceived violator, among them, possible ostracism and outright
banishment from the collective.82 To that end, prosecutors branded as
“hard” on police all too often suffer delays in gathering updates about
investigations or are completely deprived of access to information.83
Like the police, prosecutors often interact quite closely with crime
victims in the early stages of a case and may develop an allegiance to
their accounts of the event.84 Prosecutorial loyalty to victims influences
decision-making in ways contrary to the interests of criminal
defendants.85 In contrast to their relationship with victims, prosecutors
seldom become personally acquainted with the defendants in their
cases, knowing them largely through police reports and rap sheets
alone.86 Prosecutors likely have less exposure to criminal defendants
80 See Daniel Richman, Prosecutors and Their Agents, Agents and Their Prosecutors, 103
COLUM. L. REV. 749, 792 (2003); see also DAVIS, supra note 12, at 23 (noting that police officers
frequently recommend specific criminal charges to a prosecutor after making an arrest, and
prosecutors may simply follow that recommendation in certain cases); Melilli, supra note 17, at
689 (noting that due to “prolonged and recurrent contact with police officers. . . . prosecutors may
tend to regard police officers as their clients”).
81 See, e.g., Chemerinsky, supra note 78, at 305; Daniel S. Medwed, The Zeal Deal:
Prosecutorial Resistance to Post-Conviction Claims of Innocence, 84 B.U. L. REV. 125, 144-45
(2004); Abbe Smith, Can You Be a Good Person and a Good Prosecutor?, 14 GEO. J. LEGAL
ETHICS 355, 392 (2001); see also DAVIS, supra note 12, at 39-41 (discussing how prosecutors
may engage in “willful blindness” when it comes to police practices and refuse to critically
examine the steps taken in the investigation).
82 See, e.g., Susan Bandes, Loyalty to One’s Convictions: The Prosecutor and Tunnel Vision,
49 HOW. L.J. 475, 482 (2006) (“Loyalty comes with enforcement mechanisms. These include
ostracism and banishment of those who breach its rules. The costs of failing to exhibit group
loyalty can be considerable.”).
83 See, e.g., Chemerinsky, supra note 78, at 305 (“For obvious reasons, prosecutors are
reluctant to alienate the very officers that they must work with and rely on in their cases.”);
Laurie L. Levenson, The Future of State and Federal Civil Rights Prosecutions: The Lessons of
the Rodney King Trial, 41 UCLA L. REV. 509, 536 & n.150 (1994) (noting that police utilize an
array of tactics to make life difficult for a prosecutor deemed “hard on police”); Medwed, supra
note 81, at 145.
84 See Medwed, supra note 81, at 145-46.
85 See, e.g., Bandes, supra note 82, at 486; Melilli, supra note 17, at 689 (“Quite naturally,
prosecutors may develop loyalty to victims, and that loyalty may influence the prosecutors’
decisions.”).
86 See Medwed, supra note 81, at 146; see also Burke, supra note 64, at 519 (“A prosecutor
who is surrounded in her daily routine only by crime victims, police officers, and other
prosecutors might develop a deepened ‘presumption of guilt’ that can contribute to cognitive

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2205

than do the police who presumably interviewed (or tried to interview)
those parties during the investigation. Having crafted a personal link to
the victim and no corresponding connection with the accused, a
prosecutor may naturally overrate inculpatory evidence and underrate
exculpatory evidence prior to trial, to the extent that the police make
any such information available. Prosecutors may, in certain
circumstances, face external political pressure to proceed with a matter
or otherwise fear being perceived as “soft” by refusing to pursue a
tough case and take it to trial.87 Prosecutors may also, on occasion,
simply feel cynical about a particular suspect professing his innocence
because they have heard similar claims countless times before, as often
as not from clearly guilty suspects.88
For the foregoing reasons, tunnel vision often infuses a
prosecutor’s decision-making in screening a case for potential criminal
charges. The principal effect of tunnel vision in this phase is to
heighten a prosecutor’s belief in the original suspect’s guilt and
minimize any countervailing impression that someone else may have
committed the crime.89 In extreme situations, tunnel vision may prompt
prosecutors to unwittingly charge innocent suspects with crimes. Once
charges are filed, what Professor Alafair Burke calls a “sticky
presumption of guilt” often evolves, and prosecutors tend to interpret
information revealed thereafter in accord with this presumption.90
Consequently, the charging decision is a critical stage in the lifespan of
an innocent defendant’s case. A number of reforms should be
considered to limit the harm wrought by prosecutorial tunnel vision at
the charging stage.
First, requiring the police to disclose as much information as
possible to prosecutors when transferring their case files for charging

bias.”); Stanley Z. Fisher, In Search of the Virtuous Prosecutor: A Conceptual Framework, 15
AM. J. CRIM. L. 197, 208 (1988) (“[The prosecutor is regularly exposed to victims, police
officers, and others] who can graphically establish that the defendant deserves
punishment . . . . At the same time, the prosecutor is normally isolated from those—the
defendant, his family and friends, and often, his witnesses—who might arouse the prosecutor’s
empathy or stimulate concern for treating him fairly.”); Melilli, supra note 17, at 689
(“Prosecutors come to know defendants from police reports and rap sheets, and thus think of
defendants only in the context of the criminal accusations.”).
87 See Melilli, supra note 17, at 688.
88 See, e.g., Russell D. Covey, Signaling and Plea Bargaining’s Innocence Problem, 66
WASH. & LEE L. REV. 73, 92 (2009) (discussing the “signaling defect” in plea bargaining because
innocent defendants are pooled with guilty defendants also claiming innocence).
89 See, e.g., Bandes, supra note 82, at 479 (“[A] recurring theme [in wrongful convictions] is
the prosecutor’s tendency to develop a fierce loyalty to a particular version of events: the guilt of
a particular suspect or group of suspects. The loyalty is so deep it abides even when the version
of events is thoroughly discredited, or the suspect exculpated.”).
90 Burke, supra note 62, at 1605-09 (noting how selective information processing and belief
perseverance may be particularly problematic in the post-charging phase).

MEDWED.31-6

2206

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

consideration might nip one aspect of prosecutorial tunnel vision in the
bud: the fact that a prosecutor’s initial theory of guilt is ordinarily based
on incomplete information. Reviewing a complete file, as opposed to
one tailored against the person anointed the chief suspect by the police,
would aid a prosecutor in developing her theory of the case and prevent
the formulation of a presumption of guilt where flaws in the case are
readily apparent at an early stage.91 It is well within the authority of a
prosecutor’s office to demand access to such information and some
offices already do so.92 On a more general level, achieving greater
neutrality or, rather, interpersonal distance between the police and those
prosecutors entrusted with the task of reviewing police arrest files with
an eye toward possible criminal charges would be beneficial.93
Second, educating prosecutors about tunnel vision and
implementing training regimens designed to quell its influence have
their advantages as well.94 Prosecutors’ offices typically conduct inhouse continuing legal education programs already, and adding training
workshops about cognitive science is a particularly easy reform to
achieve.95 Studies suggest that urging people to articulate the specific
reasons for their position and to “counter-argue”—to take the opposite
stance from their argument as a role-play technique—can decrease the
effect of the confirmation bias and belief perseverance.96 But it is
difficult to prompt people to counter-argue. Even more, reduction, not
elimination, of tunnel vision is all that education and training offer in
the best case scenario.97
91 Id. at 1614-15 (suggesting that providing complete information might diminish the
likelihood that prosecutors will develop a theory of guilt that will trigger the host of biases, and
that, in some cases, prosecutors could even become involved in investigations prior to the
charging stage).
92 See Findley & Scott, supra note 63, at 387-88; see also Burke, supra note 62, at 1615
(“[P]olice should record, preserve, and disclose to the prosecutor all evidence collected during
their investigation, both inculpatory and exculpatory.”).
93 Findley & Scott, supra note 63, at 388 (“While it may be difficult to achieve true
neutrality, it is clear that prosecutors do manage to maintain some measure of neutrality in the
precharging context, because they do refuse to charge in a meaningful proportion of cases.”);
Melilli, supra note 17, at 673 (“And indeed, a substantial percentage of arrests result in either
declined or voluntarily aborted prosecutions.”).
94 See generally Burke, supra note 62, at 1616-18 (“Some empirical evidence suggests that
self-awareness of cognitive limitations can improve the quality of individual decision making.”);
see also Ellen S. Podgor, The Ethics and Professionalism of Prosecutors in Discretionary
Decisions, 68 FORDHAM L. REV. 1511, 1531-34 (2000) (advocating greater education both in law
schools and within prosecutors’ offices regarding discretionary decisions by prosecutors,
including the charging decision).
95 See Burke, supra note 64, at 522-23. Nevertheless, many prosecutorial educational
programs are seemingly deficient in offering training about charging decisions specifically. See
Melilli, supra note 17, at 686-87.
96 See Burke, supra note 62, at 1620; Burke, supra note 64, at 523-25; Findley & Scott, supra
note 63, at 370-71.
97 Findley & Scott, supra note 63, at 371; see also Burke, supra note 62, at 1618

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2207

Education and training therefore must be buttressed with
institutional changes. For instance, prosecutors could be urged to
promulgate written office directives to govern individual line
prosecutors in rendering charging decisions. This, in effect, would
allow prosecutors to “self-regulate” by enforcing higher charging norms
than those mandated by the codes of ethics.98 To supplement written
policies, some offices might even assign a disproportionately large
number of veteran prosecutors to their charging division to ensure
rigorous screening of police files. The New Orleans District Attorney’s
Office under Harry Connick’s leadership is perhaps the most prominent
example of an organization committed to vetting borderline cases prior
to charging and only pursuing matters in which experienced supervisors
believed they had an excellent chance of prevailing at trial.99 Whereas
instituting charging policies geared to a particular office has its
advantages, prosecutors may be reluctant to limit their charging
autonomy voluntarily.100 Office charging guidelines, where they exist,
are often devised to be as broad and flexible as possible to provide
wiggle room when defendants claim violations of them.101 More
importantly, internal guidelines alone may not be enough to counteract
tunnel vision.102 Internal regulations ideally should be coupled with
something else: the creation of a secondary review structure to take a

(“[A]lthough education about cognitive bias may hold some potential to improve prosecutorial
decision making, it is doubtful that education alone will assure prosecutorial neutrality.”).
98 See, e.g., Zacharias, supra note 60, at 762-63 (noting some of the general benefits of
internal supervision and regulation within prosecutors’ offices in limiting ethics violations).
99 See Darryl K. Brown, The Decline of Defense Counsel and the Rise of Accuracy in
Criminal Adjudication, 93 CAL. L. REV. 1585, 1619-20 (2005); Ronald Wright & Marc Miller,
The Screening/Bargaining Tradeoff, 55 STAN. L. REV. 29, 60-84 (2002). There are strong
reasons why seasoned prosecutors may do a better job than their junior colleagues in making
charging decisions. See, e.g., Rachel E. Barkow, Institutional Design and the Policing of
Prosecutors: Lessons from Administrative Law, 61 STAN. L. REV. 869, 903 (2009) (“Longevity of
service is valuable because it makes it less likely—though does not completely eliminate the
risk—that the attorney’s decision will be colored by how the decision will look to prospective
future employers.”).
100 Bennett Gershman notes that individual prosecutorial offices are free to formulate internal
guidelines to address specific types of cases, but warns that guidelines are often too general to be
of much use and that “an institutional reluctance to unduly restrict their own discretion makes it
even more unlikely that prosecutors would promulgate overly specific guidelines.” Gershman,
supra note 11, at 519-20; see also Misner, supra note 21, at 744 (“Attempts to convince
prosecutors to publish the guidelines for making prosecutorial charging decisions . . . have
generally gone unheeded. When guidelines have been drafted, they have generally been so broad
as to be of little predictive value.”).
101 See supra note 100 and accompanying text; see also Melilli, supra note 17, at 683.
102 Yaroshefsky, supra note 3, at 280-82 (discussing the Alberto Ramos case in the Bronx,
New York, and the associated discovery of widespread misconduct within the Bronx District
Attorney’s Office and inadequate internal consequences for such violations as “a glaring example
that internal controls to which disciplinary committees defer are ineffective”).

MEDWED.31-6

2208

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

“fresh look” at prosecutorial charging decisions.103
The formation of an internal review committee to evaluate
charging decisions would force the prosecutor handling the case to
communicate the precise reasons for her decision,104 a process that
could trigger the type of critical self-reflection necessary to curb tunnel
vision. Even if the charging prosecutor remained gripped by tunnel
vision despite this exercise in introspection, the members of the review
committee would likely not suffer from this malady, provided of course
that they had no previous exposure to the case.105 The practice of
seeking internal review of charging decisions is relatively common at
the federal level, with some U.S. Attorneys’ Offices requiring line
prosecutors to consult supervisors and/or committees before proceeding
with particular types of cases.106 To optimize the effectiveness of such
an internal review committee, at least one member should formally play
the role of “Devil’s Advocate” by counter-arguing and harping on the
flaws in the prosecution’s case.107 The committee admittedly may
struggle to conscript an employee who is willing and able to perform
this duty without exhibiting too much deference to their colleagues’
analyses: criticizing peers’ charging decisions day in and day out is
simply not an enviable task for a prosecutor.108 For that reason, Burke
has floated the idea of including non-prosecutors in advisory review
committees.109 A bipartisan committee of this nature might resemble
the civilian review boards that are becoming increasingly popular
mechanisms for overseeing components of police department
operations.110 Including non-prosecutors in the process would inject an
103 Findley & Scott, supra note 63, at 388-89; see also Griffin, supra note 9, at 262 (“Policies
alone cannot promote good judgment, however. That is developed through training by more
experienced prosecutors and through consultation with peers and supervisors. Accordingly, in all
matters, prosecutors should test their judgment by consulting fellow prosecutors.”).
104 See Bibas, supra note 46, at 1006 (“Simply having to explain and justify one’s decisions
disciplines prosecutors, much as writing reasoned decisions disciplines judges.”).
105 See Burke, supra note 62, at 1621 (“Another possible method to mitigate the influence of
cognitive bias on prosecutorial decision making is to involve additional, unbiased decision
makers in the process. . . . A separate attorney would be able to comment on the strength of
existing evidence against the defendant without the taint of a preexisting theory of guilt.”).
106 See Findley & Scott, supra note 63, at 388; Fitzgerald, supra note 16, at 25 (“When we
have complicated cases in my office, we have indictment committee meetings. During the
meeting, we review the case and make sure that we can prove racketeering if we are going to
charge racketeering, and look at the strengths and weaknesses of our arguments.”).
107 See Burke, supra note 62, at 1620; Findley & Scott, supra note 63, at 388-89. Susan
Bandes also recommends that “[r]eview mechanisms should exist at every level of decisionmaking,” but cautions that “review may become simply a way of reinforcing group norms.”
Bandes, supra note 82, at 493. To avoid such reinforcement, Bandes insists there must be
transparency in the form of adequate record-keeping and discovery, and that “[t]he process needs
to be explicitly structured to perform a critical role” with a “naysaying function.” Id. at 493-94.
108 See Burke, supra note 62, at 1621-22; Burke, supra note 64, at 527.
109 See Burke, supra note 62, at 1622-24.
110 Id. at 1623-24.

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2209

important element of transparency and public accountability.111
The formation of internal committees to review charging decisions
also comports with modern trends in institutional design theory.
Professor Rachel Barkow’s study of the institutional design of federal
administrative agencies spurred her to advocate greater “separation-offunctions” within U.S. Attorneys’ Offices. According to Barkow,
federal prosecutorial agencies should detach attorneys entrusted with
“investigative” tasks in a particular case from those focused on
subsequent “adjudicative” endeavors to better prevent abuses of power.
To that end, prosecutors investigating alleged criminal activity could be
barred from any involvement in the charging process.112 Barkow’s idea
has particular value in the context of federal prosecutors (the focus of
her study) who often engage in lengthy preliminary investigations of
complex criminal activities and operate in large, hierarchical
organizations. A secondary review model generally strikes me as
preferable to, and more practical than, enforcing a strict separation of
functions regime because it recognizes the efficiencies gained by having
the same prosecutor or prosecutors involved in the investigation and
charging phases of a case, especially in smaller county or local
prosecutorial agencies. On balance, internal charging review
committees embody many of the virtues of a rigid separation of
functions system—independent assessment of a potential criminal
charge by those less vested in it—without exacting too heavy an
administrative toll on the average office.
The idea of housing a charging review committee within every
prosecutorial office would encounter resistance. In particular,
establishing a committee composed in part of non-prosecutors would
face opposition from within prosecutorial ranks, impose financial
burdens, and generate qualms about conflicts of interest and
confidentiality.113 More fundamentally, the small size of many
prosecutorial agencies may make any sort of separate review committee
infeasible in those jurisdictions.114 And while larger offices may have
the resources to form internal review committees, time pressures created
by enormous caseloads in those offices accentuate the risk that the
review process will be largely ceremonial.115 But, regardless of the
precise makeup of internal charging review committees and the
111 See generally Bandes, supra note 82, at 493-94 (describing the important role transparency
plays in the charging process); Findley & Scott, supra note 63, at 391 (same).
112 See Barkow, supra note 99, at 898.
113 See Burke, supra note 62, at 1623; Burke, supra note 64, at 527.
114 Barkow acknowledges that “a panel of adjudicative decision makers is preferred” to
counter individual biases, but suggests it might not be feasible in many offices. Barkow, supra
note 99, at 904.
115 See Brown, supra note 99, at 1620; Melilli, supra note 17, at 683.

MEDWED.31-6

2210

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

practical barriers to their creation, the general concept has much to
recommend it as a tool to alleviate the effects of tunnel vision, and such
committees should be implemented where possible.
It would be unwise, however, to compel internal review
committees to evaluate each and every charging decision. Many routine
cases in our burgeoning, unrelenting criminal justice system may not
merit extensive outside evaluation.116 In recognition of the desire for
efficiency and the existence of resource constraints (both time and
money),117 committee review of charging decisions should be confined
to those cases most deserving of scrutiny: in other words, those cases
where the risk of a wrongful conviction is most pronounced.118 It is not
only high-profile or complicated cases that carry such a risk,119 but also
those imbued with certain characteristics found to prevail in a large
swath of wrongful convictions.120
Indeed, empirical studies of post-conviction exonerations of
innocent prisoners have isolated a distinct subset of factors that
contribute most dramatically to wrongful convictions: eyewitness
misidentifications, false confessions, jailhouse informants, police and
prosecutorial misconduct, use of dubious forensic science, and
ineffective
assistance
of
defense
counsel.121
Eyewitness
misidentification stands out as the most common variable in the
conviction of the innocent, as highlighted by the Innocence Project’s
2010 report concluding that such errors occurred at trial in 76% of the
first two hundred and fifty exonerations achieved through postconviction DNA testing.122
What is more, 53% of those
116 Cookie-cutter, “buy-and-bust” drug cases where an undercover officer has purchased drugs
from an oblivious seller and the “buy money” is found in the seller’s possession come to mind as
paradigmatic examples of cases that fail to cry out for committee review.
117 See, e.g., Melilli, supra note 17, at 683 (observing that charging decisions “must often be
made spontaneously and instinctively with infrequent opportunities for serious internal review”).
118 See Burke, supra note 64, at 526 (“Fresh looks would appear to be particularly helpful in
cases where some of the government’s original evidence against a defendant has been
undermined; a new lawyer could review the case considering only the remaining evidence,
untainted by the lingering effects of belief perseverance.”).
119 See supra notes 57-59 and accompanying text (discussing Professor Mosteller’s advocacy
of a prosecutorial duty to investigate “high-profile” or “problematic” cases before filing charges);
see also Burke, supra note 64, at 526 (“Offices with sufficient resources could create a formal
layer of internal review, at least in some limited categories of high-stakes cases, such as death
penalty cases, other major crimes, or post-conviction claims of innocence.”).
120 Indeed, a failing with many of the pre-existing prosecutorial charging review committees
and scholarly calls for further enactment is their almost exclusive emphasis on complex cases.
See, e.g., Fitzgerald, supra note 16, at 25; Griffin, supra note 9, at 293 (noting how prosecutorial
agencies frequently have departmental review processes for charging decisions related to capital
punishment).
121 See, e.g., BARRY C. SCHECK ET AL., ACTUAL INNOCENCE: WHEN JUSTICE GOES WRONG
AND HOW TO MAKE IT RIGHT (2001) (discussing a number of systemic flaws that generate
wrongful convictions, and listing reforms to protect the innocent).
122 See INNOCENCE PROJECT, 250 EXONERATED: TOO MANY WRONGLY CONVICTED 22-23,

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2211

misidentifications involved witnesses and perpetrators of different
races, so-called “cross-racial” identifications.123 False confessions by
innocent suspects, especially juveniles and the developmentallydisabled, also take place with surprising frequency.124 Furthermore,
informants played a role in 19% of the wrongful convictions cited in the
Innocence Project’s report.125 Prosecutorial reliance at trial on forensic
evidence based on unsound scientific principles cropped up as a central
factor in 52% of the cases studied.126
Although some of the chief factors that generate wrongful
convictions—such as most forms of police and prosecutorial
misconduct and ineffective assistance of defense counsel—are not
readily identifiable prior to trial, many are. For instance, pretrial
prosecutors stand in a reasonably solid position to weigh the accuracy of
an eyewitness’s identification. They can gauge the conditions under
which the crime occurred, discern whether “cross-racial” identification
is an issue, and examine the photo array and physical lineup procedures
to determine whether there was a chance of suggestiveness.127
Likewise, prosecutors can rather easily pinpoint whether the case
against a suspect rests on forensic evidence and, if so, whether that
evidence derives from a forensic technique susceptible to charges of
inaccuracy, such as the essentially discredited field of “hair
microscopy” in which forensic scientists sought to conclude through
visual inspection whether a hair follicle retrieved from the crime scene
matched that obtained from a suspect.128 Ascertaining the existence of a
false confession poses more of a challenge, but, at a minimum,
prosecutors might critically examine statements by juvenile and
mentally-deficient suspects prior to trial. And, to be sure, prosecutors
are ideally situated well before trial to assess the credibility of
informants that they intend to present as witnesses.129
On the whole, prosecutorial committees assigned to review
charging decisions should steer their efforts primarily in the direction of
www.innocenceproject.org/docs/InnocenceProject_250.pdf [hereinafter INNOCENCE PROJECT]
(last visited June 25, 2010).
123 Id. at 24-25.
124 See Gross et al., supra note 2, at 544-46.
125 INNOCENCE PROJECT, supra note 122, at 38-39.
126 Id. at 28-29.
127 See, e.g., Daniel S. Medwed, Anatomy of a Wrongful Conviction: Theoretical Implications
and Practical Solutions, 51 VILL. L. REV. 337, 358-364 (2006).
128 See INNOCENCE PROJECT, supra note 122, at 30 (mentioning that twenty-one percent of the
exonerations in their study involved the use of hair microscopy at trial).
129 For example, in the Los Angeles County District Attorney’s Office, an in-house jailhouse
informant committee must grant prior approval of an informant’s appearance at trial before a
prosecutor may use that person as a witness. See Myrna S. Raeder, See No Evil: Wrongful
Convictions and the Prosecutorial Ethics of Offering Testimony by Jailhouse Informants and
Dishonest Experts, 76 FORDHAM L. REV. 1413, 1449 (2007).

MEDWED.31-6

2212

8/9/2010 7:44:30 PM

CARDOZO LAW REVIEW

[Vol. 31:6

those matters containing indicia of possible eyewitness
misidentifications, false confessions, and/or overreliance on shaky
forensic findings. To the extent that this cohort of cases might remain
too vast and thus impractical for review committees to tackle, I would
suggest targeting charging decisions predicated on a “single
eyewitness” and little, if any, other evidence. Those cases present the
greatest hazard because of the high rate of identification error and,
paradoxically, the high esteem with which jurors tend to hold
eyewitness testimony.130 A handful of prosecutors’ offices have
launched “single eyewitness” review committees.131 One such
committee embedded within the Nassau County District Attorney’s
Office in New York State at one point appraised roughly ten cases per
year and dismissed two of them on average.132 Would any innocent
defendants have been convicted if charges had been filed in the cases
jettisoned by the review committee in Nassau County? It is impossible
to know for certain, but I strongly suspect the answer is yes.
Tunnel vision is an innate human trait and, as such, incurable.
Still, its worst symptoms can receive treatment in the context of
prosecutors choosing whether to charge a particular suspect with a
crime. The promulgation of a secondary review committee within every
prosecutor’s office to take a fresh look at charging decisions in cases
possessing the hallmarks of wrongful convictions would likely lead to
the dismissal of especially weak cases prior to the submission of formal
charges. This would go a long way toward making the impact of tunnel
vision less fatal for the innocent.
CONCLUSION
At its core, the charging decision is the tipping point for a criminal
case. If the prosecutor comes out against charging, then the case
virtually disappears with little harm. If the prosecutor sides in favor of
filing charges, the government’s efforts tilt toward developing a case for
trial. Once the wheels of a criminal case are set in motion toward trial,
the chance of a wrongful conviction increases significantly. That is
130 For example, in an experiment involving a robbery trial, a jury found the defendant guilty
eighteen percent of the time where the prosecution’s case lacked an eyewitness. ELIZABETH F.
LOFTUS, EYEWITNESS TESTIMONY 9-10 (1979). The addition of a single eyewitness to those
experimental cases boosted the conviction rate to seventy-two percent. Id.
131 Several chief prosecutors have embarked on such a path. See, e.g., Sean Gardiner, For
Them, No Justice; Bad Convictions Put 13 Men in Prison. Persistence-and Luck-Got Them Out,
NEWSDAY, Dec. 8, 2002, at A03; Robin Topping, Panel Puts Justice Before Prosecution,
NEWSDAY, Jan. 8, 2003, at A21.
132 See Topping, supra note 131.

MEDWED.31-6

2010]

8/9/2010 7:44:30 PM

EMOTIONALLY CHARGED

2213

exactly why reforms to the rules and practices surrounding the charging
decision deserve careful consideration: to stop the wheels from turning
at all in weak cases.