Holmes, Watson and the Riddle of Harmless Error ©
By Joseph Baxter, Certified Appellate Specialist
By Joseph Baxter, Certified Appellate Specialist
Few principles of American law are more well settled than the principles underlying the division of labor between our trial and appellate courts: juries and trial judges are ultimate arbiters of the facts while appellate courts are ultimate arbiters of the law. Yet, this clarity breaks down when a legal error is established and the appellate court must decide whether the error actually prejudiced the ultimate outcome. We have what Justice Roger Traynor called “that most pervasive and elusive of all problems in an appellate court, the riddle of harmless error.”
The earliest decisions declared the riddle insoluble. Nineteenth century British and American courts admitted they were incapable of determining whether an error affected the judgment. It was a question deemed too difficult to answer “in the light or the dark of the trial.”
The judicial reflex eventually degenerated into absurdity, as British and American appellate courts became impregnable citadels of technicality. No error was lightly forgiven. Thus, in 1880, the Supreme Court of California reversed a larceny conviction because the indictment charged the crime as “larcey.” As Justice Traynor quipped, larcey could not “now be laced up with an n.” “New trials were ordered at the drop of a hat or a consonant that was needed to split a hair.”
The earliest decisions declared the riddle insoluble. Nineteenth century British and American courts admitted they were incapable of determining whether an error affected the judgment. It was a question deemed too difficult to answer “in the light or the dark of the trial.”
The judicial reflex eventually degenerated into absurdity, as British and American appellate courts became impregnable citadels of technicality. No error was lightly forgiven. Thus, in 1880, the Supreme Court of California reversed a larceny conviction because the indictment charged the crime as “larcey.” As Justice Traynor quipped, larcey could not “now be laced up with an n.” “New trials were ordered at the drop of a hat or a consonant that was needed to split a hair.”
THE MISCARRIAGE OF JUSTICE TEST
Statutory and constitutional changes were ultimately implemented to deal with this absurdity. In California a 1911 constitutional amendment prohibited the reversal of a criminal conviction for error “unless, after an examination of the entire cause including the evidence, the court shall be of the opinion that the error complained of resulted in a miscarriage of justice.” The amendment was soon extended to apply to civil judgments.
The ballot argument for the adoption of the 1911 constitutional change sounds surprisingly modern.
The amendment was “designed to meet the ground of common complaint that criminals escape justice through technicalities.” The courts have gone to “absurd lengths . . . in the reversal of cases for immaterial errors . . . .” “The amendment . . . is designed to render it impossible for the higher courts to reverse the judgments of our trial courts in criminal cases for unimportant errors.”
Statutory and constitutional changes were ultimately implemented to deal with this absurdity. In California a 1911 constitutional amendment prohibited the reversal of a criminal conviction for error “unless, after an examination of the entire cause including the evidence, the court shall be of the opinion that the error complained of resulted in a miscarriage of justice.” The amendment was soon extended to apply to civil judgments.
The ballot argument for the adoption of the 1911 constitutional change sounds surprisingly modern.
The amendment was “designed to meet the ground of common complaint that criminals escape justice through technicalities.” The courts have gone to “absurd lengths . . . in the reversal of cases for immaterial errors . . . .” “The amendment . . . is designed to render it impossible for the higher courts to reverse the judgments of our trial courts in criminal cases for unimportant errors.”
THE “MAY WELL HAVE BEEN SUBSTANTIALLY INJURED” TEST
A California constitutional amendments (section 4½ of Article VI) quickly brought about a sea change in appellate law. The first generation of decisions ruled that reversal is only required where the appellant “may well have been substantially injured by the error of which he complains.” People v. Watts (1926) 198 Cal. 776, 792.
A second generation of decisions, however, began to produce variations on this theme. Decisions increasingly favored a double negative test. Reversal was required where “a different verdict would not have been improbable had the error not occurred.”
THE “REASONABLE PROBABILITY” TEST
In 1956, the California Supreme Court attempted to harmonize this “varying language,” but with yet another variation. People v. Watson, perhaps the most widely cited case in California history, arose out of a trial of a young Army Corporal for the murder of his wife. While the wife’s body was found in the bathtub, bruises and lacerations on her body established multiple traumatic injuries as the cause of death. The jury found Watson guilty of second degree murder.
Corporal Watson admitted that he and his wife had an argument the night before. According to the court, moreover, the undisputed decomposition evidence established that the death occurred “during the hours that defendant admittedly was in the apartment,” on the day of her death.
The six foot six Army Corporal was entitled – perhaps even required – to wear his uniform while testifying. The government, nonetheless, taunted him for doing exactly that, cross-examining him about his alleged efforts to stretch his height to obtain a discharge, and his application for an early discharge. He was grilled about marks on his apartment walls indicating his height change and his gymnasium exercises. In the Supreme Court’s view, the prosecution’s attempts to degrade Watson for wearing his uniform were clearly improper.
This was not the only error committed below. While the trial court instructed on reasonable doubt, it refused to instruct the jury that a case resting entirely on links of circumstantial evidence also required proof of each link beyond a reasonable doubt. According to the Supreme Court, this was also clearly improper since the case against Corporal Watson was entirely circumstantial.
Acknowledging the errors below, the Supreme Court proceeded to an evaluation of whether the errors were “harmless.” The Court recognized the “varying language” of California tests for prejudicial error, noting both the Watts test (“may well have been substantially injured”) and the “not improbable” test (“A different result would not have been improbable had the error not occurred.”). In the Court’s view these variations of language could be crystallized into one overarching test:
That a “miscarriage of justice” should be declared only when the court, . . . is of the “opinion” that it is reasonably probable that a result more favorable to the appealing party would have been reached in the absence of the error. Phrasing the test in this language avoids any complexity which may be said to result from the language employed in the double negative approach . . . .
In 1956, the California Supreme Court attempted to harmonize this “varying language,” but with yet another variation. People v. Watson, perhaps the most widely cited case in California history, arose out of a trial of a young Army Corporal for the murder of his wife. While the wife’s body was found in the bathtub, bruises and lacerations on her body established multiple traumatic injuries as the cause of death. The jury found Watson guilty of second degree murder.
Corporal Watson admitted that he and his wife had an argument the night before. According to the court, moreover, the undisputed decomposition evidence established that the death occurred “during the hours that defendant admittedly was in the apartment,” on the day of her death.
The six foot six Army Corporal was entitled – perhaps even required – to wear his uniform while testifying. The government, nonetheless, taunted him for doing exactly that, cross-examining him about his alleged efforts to stretch his height to obtain a discharge, and his application for an early discharge. He was grilled about marks on his apartment walls indicating his height change and his gymnasium exercises. In the Supreme Court’s view, the prosecution’s attempts to degrade Watson for wearing his uniform were clearly improper.
This was not the only error committed below. While the trial court instructed on reasonable doubt, it refused to instruct the jury that a case resting entirely on links of circumstantial evidence also required proof of each link beyond a reasonable doubt. According to the Supreme Court, this was also clearly improper since the case against Corporal Watson was entirely circumstantial.
Acknowledging the errors below, the Supreme Court proceeded to an evaluation of whether the errors were “harmless.” The Court recognized the “varying language” of California tests for prejudicial error, noting both the Watts test (“may well have been substantially injured”) and the “not improbable” test (“A different result would not have been improbable had the error not occurred.”). In the Court’s view these variations of language could be crystallized into one overarching test:
That a “miscarriage of justice” should be declared only when the court, . . . is of the “opinion” that it is reasonably probable that a result more favorable to the appealing party would have been reached in the absence of the error. Phrasing the test in this language avoids any complexity which may be said to result from the language employed in the double negative approach . . . .
In the court’s view, the new test was “essentially the same” as the prior tests:
We are of the view, however, that the test as above stated does not constitute a departure from the tests heretofore applied . . . . [T]he tests, as variously stated, are not fundamentally different but, on the contrary, are essentially the same. Nevertheless, the test, as stated in any of the several ways, must necessarily be based upon reasonable probabilities rather than upon mere possibilities . . . .
Applying this newly minted test, the Court concluded that the degrading cross-examination of Corporal Watson’s was harmless error since the defendant was allowed to explain his actions. The Court also held that the trial court’s refusal to give a required circumstantial evidence/reasonable doubt instruction was not prejudicial since the uncontested decomposition evidence established that the death occurred during the hours defendant was in the apartment.
THE “ESSENTIALLY THE SAME” TESTS
While Watson may well be the most widely cited case in California history, the decision was flawed on many levels. The court adopted a vague, and almost meaningless abstract language – the “reasonable probability” test – to convey the meaning of other standards, i.e., the Watts “may well have been substantially prejudiced” test and the Putnam “not improbable” test, both of which conveyed a greater level of meaning. At best, the court accomplished nothing. At worst, it accomplished obfuscation.
Even after the closest of readings, it is not at all clear where Watson draws its benchmark. The referenced Watts “may well have been substantially prejudiced” test suggests that substantial prejudice must be a real possibility, as opposed to an abstract possibility. Thus, if we say that “it may well rain tomorrow” we simply mean that rain is a real possibility. By itself, however, the Watts “may well have” language does not suggest a specific degree of probability. Thus, if only one of five possible individuals can appear through a door, and one of them definitely will, one can properly say that any one of them “may well” appear through the door. By the same token, if one in thirty possible individuals can appear through a door, and only one of them definitely will, one can also say that any of them “may well” appear through the door.
The “not improbable” test also does not suggest a specific degree of probability. Since only remote possibilities are “improbable,” the “not improbable” test essentially suggests that prejudice must be something other than a remote possibility.
Arguably, these pre-Watson tests suggest a low – (5 percent?) – level of probability of prejudice required for reversal. If so, this would appear to be consistent with the ballot arguments for the adoption of the 1911 amendment, which was aimed at preventing reversals for “technicalities.”
As a term, however, “reasonable probability,” does not really convey any particular meaning. In the strict sense of the word, an actual probability is “reasonable” since reason is always required to deduce them. As Amir Aczel, Ph.D. states in his classic work, Chance, a “probability is a quantitative measure of the likelihood of a given event.” By this view, however, the Watson “reasonable probability” is an oxymoron since Watson’s “reasonable” is expressly non-quantitative. A ten percent probability, a thirty percent probability, and an eighty percent probability are lumped together. All are “reasonable” so long as they are real.
THE “ESSENTIALLY THE SAME” TESTS
While Watson may well be the most widely cited case in California history, the decision was flawed on many levels. The court adopted a vague, and almost meaningless abstract language – the “reasonable probability” test – to convey the meaning of other standards, i.e., the Watts “may well have been substantially prejudiced” test and the Putnam “not improbable” test, both of which conveyed a greater level of meaning. At best, the court accomplished nothing. At worst, it accomplished obfuscation.
Even after the closest of readings, it is not at all clear where Watson draws its benchmark. The referenced Watts “may well have been substantially prejudiced” test suggests that substantial prejudice must be a real possibility, as opposed to an abstract possibility. Thus, if we say that “it may well rain tomorrow” we simply mean that rain is a real possibility. By itself, however, the Watts “may well have” language does not suggest a specific degree of probability. Thus, if only one of five possible individuals can appear through a door, and one of them definitely will, one can properly say that any one of them “may well” appear through the door. By the same token, if one in thirty possible individuals can appear through a door, and only one of them definitely will, one can also say that any of them “may well” appear through the door.
The “not improbable” test also does not suggest a specific degree of probability. Since only remote possibilities are “improbable,” the “not improbable” test essentially suggests that prejudice must be something other than a remote possibility.
Arguably, these pre-Watson tests suggest a low – (5 percent?) – level of probability of prejudice required for reversal. If so, this would appear to be consistent with the ballot arguments for the adoption of the 1911 amendment, which was aimed at preventing reversals for “technicalities.”
As a term, however, “reasonable probability,” does not really convey any particular meaning. In the strict sense of the word, an actual probability is “reasonable” since reason is always required to deduce them. As Amir Aczel, Ph.D. states in his classic work, Chance, a “probability is a quantitative measure of the likelihood of a given event.” By this view, however, the Watson “reasonable probability” is an oxymoron since Watson’s “reasonable” is expressly non-quantitative. A ten percent probability, a thirty percent probability, and an eighty percent probability are lumped together. All are “reasonable” so long as they are real.
THE WATSONTHE WATSON DISSENTS
Two Justices of the California Supreme Court issued stinging dissents from the majority decision in Watson. Dissenting Justice Shauer noted that “[m]any of the suspicious circumstances are susceptible to innocent construction” and that “resolution of conflicting inferences” were for the jury. Given that “the scales of proof are so delicately balanced” it should be resolved by a jury “which has not been prejudiced.”
Justice Carter noted that three Justices of the Court of Appeal had reversed the conviction because they could not “say with conviction that, in the absence of the errors complained of, a different verdict would have been improbable . . . .” In Justice Carters’ view, the philosophy of the individual members of this court” was at play, rather than any real attempt at interpreting the California Constitution. In his view, the majority opinion was simply “double-talk”:
A reading of the majority opinion here demonstrates that whenever the majority desires to invoke section 4 ½ of article VI for the purpose of affirming a judgment of conviction, it simply resorts to double-talk and finally arrives at the conclusion that the judgment should be affirmed. This is precisely what has happened in the case at bar.
Two Justices of the California Supreme Court issued stinging dissents from the majority decision in Watson. Dissenting Justice Shauer noted that “[m]any of the suspicious circumstances are susceptible to innocent construction” and that “resolution of conflicting inferences” were for the jury. Given that “the scales of proof are so delicately balanced” it should be resolved by a jury “which has not been prejudiced.”
Justice Carter noted that three Justices of the Court of Appeal had reversed the conviction because they could not “say with conviction that, in the absence of the errors complained of, a different verdict would have been improbable . . . .” In Justice Carters’ view, the philosophy of the individual members of this court” was at play, rather than any real attempt at interpreting the California Constitution. In his view, the majority opinion was simply “double-talk”:
A reading of the majority opinion here demonstrates that whenever the majority desires to invoke section 4 ½ of article VI for the purpose of affirming a judgment of conviction, it simply resorts to double-talk and finally arrives at the conclusion that the judgment should be affirmed. This is precisely what has happened in the case at bar.
Justice Carter’s criticism certainly appears to hit its mark. The Watson court doesn’t really apply its newly minted “reasonable probability” test. Thus, the court simply announces its conclusion that the degrading cross-examination really wasn’t that prejudicial. The court essentially concluded that the cross-examination was harmless since the defendant got an opportunity to explain his actions. The Court’s analysis, however, misses the point. No matter what Corporal Watson might have to say about the evidence tending to show his efforts at gaining a discharge, the improper cross-examination had the effect of turning a six foot six patriot into a moral midget.
THE SURPRISING NATURE OF PROBABILITIES
The Watson decision also ignores the fact that ordinary human perceptions of probabilities are often unreliable. Consider the following examples:
1. A repeated coin toss will not necessarily produce an equal number of heads and tails – the probable result – until many tosses of the coin.
2. How many people do you need to have in a room before there is a 50 percent probability that two of them will have the same birthday? The common expectation is a number larger than a hundred. Surprisingly, it only takes 23.
3. Assume you are betting red or black on a roulette wheel with a 47 percent probability of success, i.e., if you bet $10 you have a 47 percent chance of collecting $20. If you have a thousand dollar cap on your betting expenses, and are willing to gamble it all, is it wiser to bet it all on the first spin, or is it wiser to place many small bets, e.g., $50 a bet? Or do both strategies yield the same likelihood of success? Bet it all on the first spin, and leave. Your first bet has a 47 percent probability of success. If you bet fifty dollars every time, you have only an 11 percent chance of success.
The Watson decision also ignores the fact that ordinary human perceptions of probabilities are often unreliable. Consider the following examples:
1. A repeated coin toss will not necessarily produce an equal number of heads and tails – the probable result – until many tosses of the coin.
2. How many people do you need to have in a room before there is a 50 percent probability that two of them will have the same birthday? The common expectation is a number larger than a hundred. Surprisingly, it only takes 23.
3. Assume you are betting red or black on a roulette wheel with a 47 percent probability of success, i.e., if you bet $10 you have a 47 percent chance of collecting $20. If you have a thousand dollar cap on your betting expenses, and are willing to gamble it all, is it wiser to bet it all on the first spin, or is it wiser to place many small bets, e.g., $50 a bet? Or do both strategies yield the same likelihood of success? Bet it all on the first spin, and leave. Your first bet has a 47 percent probability of success. If you bet fifty dollars every time, you have only an 11 percent chance of success.
4. Assume that one out of every thousand people has a particular disease and assume that the test for the disease has a 95 percent reliability, meaning that if the person tested is sick, there is a 95 percent probability that the test will indicate so.” Conversely, there is a five percent probability that healthy people will be reported as sick. If a person is reported as sick, what is the probability that the person is actually sick? One might assume a 95 percent probability. The actual probability, however, is less than two percent.
5. According to a study published in Nature, the United Kingdom Meteorological Office 24-hour forecast of rain achieves 83 percent accuracy, while the average probability of rain on an hourly time scale is eight percent. If, after an extended period of time, an individual has taken 1,000 one-hour walks after hearing forecasts of rain, how accurate were the forecasts? One might assume that he would have gotten wet 83 percent of the time. However, the correct answer is 30 percent.
5. According to a study published in Nature, the United Kingdom Meteorological Office 24-hour forecast of rain achieves 83 percent accuracy, while the average probability of rain on an hourly time scale is eight percent. If, after an extended period of time, an individual has taken 1,000 one-hour walks after hearing forecasts of rain, how accurate were the forecasts? One might assume that he would have gotten wet 83 percent of the time. However, the correct answer is 30 percent.
6. Our probability evaluations are often dictated or influenced by our emotions. Imagine that you are a New York cab driver with an income that varies from day to day, depending on the weather, the level of tipping, the number of fares, etc. What probability strategy do you use to make the most money while working the fewest hours? One approach would be to put in more hours on good income days, while quitting early on bad income days. If you follow this strategy your net income would increase by 15 percent. Yet, according to a study by economist Colin Camerer of the California Institute of Technology, New York cab drivers adopt the most irrational strategy: they quit early on high income days and work longer on low income days.
As these examples illustrate, probabilities are subtle and often counter-intuitive, and our probability evaluations are not always rational. According to Richard Restak, MD, “the brain doesn’t do a very good job when it comes to estimating probability.” Watson.” Watson, however, incorrectly assumes that our perceptions of probabilities are always reliable. Perhaps this is not surprising since Watson.” Watson, however, incorrectly assumes that our perceptions of probabilities are always reliable. Perhaps this is not surprising since Watson entirely ignores the difficult problem of quantifying probability.
THE FIFTY-YEAR FOG
Whatever elusive meaning “reasonable probability” may have had when Watson was decided, California appellate courts quickly lost grasp of it. “Reasonable probability” turned into a mantra with absolutely no examination of what the mantra means. Thus, thousands of California appellate decisions recite a nearly identical claim: reversal is only required “where it is reasonably probable the error affected the outcome.” The ordinary connotations associated with this choice of language appear to suggest that “reasonably probable” means at least 50 percent probability. People v. Watson, however, doesn’t require proof of 50 percent probability. According to the Watson court, a “reasonable probability” is, instead, a probability that is larger than a mere possibility.
Some decisions have suggested that “reasonable probability” means “more likely than not.” The California Supreme Court, however, has declared that “reasonable probability” does not mean “more likely than not.”
The Watson “reasonable probability” test is also sometimes confused with the Strickland v. Washington (1984) 466 U.S. 668 “reasonable probability” test in which “reasonable probability” is defined as a probability that “undermines confidence in the outcome.” Some decisions, even by the California Supreme Court, have suggested that this Strickland “reasonably probable” means “probable,” which it does not.
Even esteemed legal authorities can lose their way in the thick Watson fog. Roger Traynor, the dean of harmless error analysis, got the Watson standard backwards in his classic essay. Justice Epstein incorrectly suggests that “reasonably probable,” means “probable.”
As these examples illustrate, probabilities are subtle and often counter-intuitive, and our probability evaluations are not always rational. According to Richard Restak, MD, “the brain doesn’t do a very good job when it comes to estimating probability.” Watson.” Watson, however, incorrectly assumes that our perceptions of probabilities are always reliable. Perhaps this is not surprising since Watson.” Watson, however, incorrectly assumes that our perceptions of probabilities are always reliable. Perhaps this is not surprising since Watson entirely ignores the difficult problem of quantifying probability.
THE FIFTY-YEAR FOG
Whatever elusive meaning “reasonable probability” may have had when Watson was decided, California appellate courts quickly lost grasp of it. “Reasonable probability” turned into a mantra with absolutely no examination of what the mantra means. Thus, thousands of California appellate decisions recite a nearly identical claim: reversal is only required “where it is reasonably probable the error affected the outcome.” The ordinary connotations associated with this choice of language appear to suggest that “reasonably probable” means at least 50 percent probability. People v. Watson, however, doesn’t require proof of 50 percent probability. According to the Watson court, a “reasonable probability” is, instead, a probability that is larger than a mere possibility.
Some decisions have suggested that “reasonable probability” means “more likely than not.” The California Supreme Court, however, has declared that “reasonable probability” does not mean “more likely than not.”
The Watson “reasonable probability” test is also sometimes confused with the Strickland v. Washington (1984) 466 U.S. 668 “reasonable probability” test in which “reasonable probability” is defined as a probability that “undermines confidence in the outcome.” Some decisions, even by the California Supreme Court, have suggested that this Strickland “reasonably probable” means “probable,” which it does not.
Even esteemed legal authorities can lose their way in the thick Watson fog. Roger Traynor, the dean of harmless error analysis, got the Watson standard backwards in his classic essay. Justice Epstein incorrectly suggests that “reasonably probable,” means “probable.”
THE FATHER OF WATSON STILL SURVIVES UNDER HIS SON’S NAME
Watson, after all, was the son of Watts. It should therefore be no surprise that a number of post-Watson prejudicial error opinions incorporate Watts-style “may well have been substantially injured” language. Thus, in People v. Bell, Justice Corrigan (now a member of the California Supreme Court) enforced the Watts standard while equating it with the Watson standard:
“Given the state of the evidence we conclude that Bell ‘may well have been substantially injured by the error of which he complains.’ (People v. Watts (1926) 198 Cal. 776, 793) It is reasonably probable he might have achieved a more favorable result had the instruction not been given. (People v. Hannon, supra, 19 Cal.3d at p. 603; People v. Watson (1956) 46 Cal.2d 818, 836.”
Watson, after all, was the son of Watts. It should therefore be no surprise that a number of post-Watson prejudicial error opinions incorporate Watts-style “may well have been substantially injured” language. Thus, in People v. Bell, Justice Corrigan (now a member of the California Supreme Court) enforced the Watts standard while equating it with the Watson standard:
“Given the state of the evidence we conclude that Bell ‘may well have been substantially injured by the error of which he complains.’ (People v. Watts (1926) 198 Cal. 776, 793) It is reasonably probable he might have achieved a more favorable result had the instruction not been given. (People v. Hannon, supra, 19 Cal.3d at p. 603; People v. Watson (1956) 46 Cal.2d 818, 836.”
Many decisions, moreover, employ the Watts standard without actually citing Watts. Thus, in People v. Hannon, the Supreme Court of California held,
“This was a close case and it appears reasonably probable that a verdict more favorable to the defendant might have resulted if the error had not occurred. (People v. Watson (1956) 46 Cal.2d 818, 836) Accordingly, the judgment must be reversed.”
“This was a close case and it appears reasonably probable that a verdict more favorable to the defendant might have resulted if the error had not occurred. (People v. Watson (1956) 46 Cal.2d 818, 836) Accordingly, the judgment must be reversed.”
EMERGING FROM THE WATSON FOG
In 1994 the California Supreme Court adopted a new and better articulation of the rule of the Watson case. In College Hospital, Inc. v. Superior Court, the California high court held that a “reasonable probability exists when there is merely a reasonable chance, more than an abstract possibility” of a different outcome. Eight decisions by the California Supreme Court and the California Courts of Appeal have re-affirmed the new articulation of the Watson rule.
In 1994 the California Supreme Court adopted a new and better articulation of the rule of the Watson case. In College Hospital, Inc. v. Superior Court, the California high court held that a “reasonable probability exists when there is merely a reasonable chance, more than an abstract possibility” of a different outcome. Eight decisions by the California Supreme Court and the California Courts of Appeal have re-affirmed the new articulation of the Watson rule.
OLD HABITS
Old habits, nonetheless, die slowly. While nine published California decisions cite or rely upon the rule that “reasonable probability” means a “reasonable chance, more than an abstract possibility,” many hundreds of California appellate decisions continue to ignore the rule and only apply the Watson “reasonable probability” mantra. Thus, in the two years since its Cassim v. Allstate Insurance Company decision expressly reminded lower appellate courts that “reasonable probability” means “reasonable chance,” only one of the high court decisions cite or rely upon the new rule, while seventeen opinions cite or rely upon People v. Watson without paying heed to the new rule.
The emerging pattern of California decisions is, nonetheless, absolutely clear. The College Hospital rule is slowly being applied to civil cases. Yet no California Supreme Court decision has applied it to a criminal judgment, and only one published Court of Appeal decision has applied it to a criminal judgment.
This unequal application of the California harmless error rule cannot continue. The civil and criminal harmless error tests for non-constitutional error are the same. Cassim v. Allstate Ins. Co. (2004) 33 Cal.4th 780,801; Sonle v. BM Corp. (1994) Cal.4th 548, 576 Both arise from the same state constitutional provision. It can hardly be argued that error is more serious when money or property is at stake than when life and liberty are at stake.
THE RIDDLE OF HARMLESS ERROR
People v. Watson is the most widely cited appellate decision in California history. Yet it has also been the least understood. How the most cited decision in California appellate history can be the least understood is yet another riddle of harmless error.
Old habits, nonetheless, die slowly. While nine published California decisions cite or rely upon the rule that “reasonable probability” means a “reasonable chance, more than an abstract possibility,” many hundreds of California appellate decisions continue to ignore the rule and only apply the Watson “reasonable probability” mantra. Thus, in the two years since its Cassim v. Allstate Insurance Company decision expressly reminded lower appellate courts that “reasonable probability” means “reasonable chance,” only one of the high court decisions cite or rely upon the new rule, while seventeen opinions cite or rely upon People v. Watson without paying heed to the new rule.
The emerging pattern of California decisions is, nonetheless, absolutely clear. The College Hospital rule is slowly being applied to civil cases. Yet no California Supreme Court decision has applied it to a criminal judgment, and only one published Court of Appeal decision has applied it to a criminal judgment.
This unequal application of the California harmless error rule cannot continue. The civil and criminal harmless error tests for non-constitutional error are the same. Cassim v. Allstate Ins. Co. (2004) 33 Cal.4th 780,801; Sonle v. BM Corp. (1994) Cal.4th 548, 576 Both arise from the same state constitutional provision. It can hardly be argued that error is more serious when money or property is at stake than when life and liberty are at stake.
THE RIDDLE OF HARMLESS ERROR
People v. Watson is the most widely cited appellate decision in California history. Yet it has also been the least understood. How the most cited decision in California appellate history can be the least understood is yet another riddle of harmless error.
One explanation is that appellate judges actually like the Watson ambiguity since it allows them to reach what they believe is the right result without unnecessary effort. The truth is that the intense workload of our modern appellate courts does not allow our appellate judges the leisure of reading lengthy records. Yet one may rightly ask whether a court that does not read the record should weigh probabilities expressly based on a review of the record. To use Justice Scalia’s choice of words, when a jury trial involves error, upholding a conviction without reading the record may be akin to “operating on the spinal column of American democracy” while the lights are out.
Roger Traynor was himself of the view that judicial bias can invite all too easy affirmance in criminal appeals.
Appellate judges, persuaded by the record that the defendant committed some crime, are often reluctant to open the way to a new trial, given not only the risk of draining judicial resources but also the risk that a guilty defendant may go free.
Roger Traynor was himself of the view that judicial bias can invite all too easy affirmance in criminal appeals.
Appellate judges, persuaded by the record that the defendant committed some crime, are often reluctant to open the way to a new trial, given not only the risk of draining judicial resources but also the risk that a guilty defendant may go free.
Traynor’s view may explain the current dichotomization of harmless error in California: with rare exception, the College Hospital “reasonable chance” test is only being applied in civil appeals.
THE PATH OF THE LAW
Roger Traynor strongly believed that the Watson “reasonable probability” test was both too nebulous and too error-affirming. His classic essay concluded with an appeal for a new standard: “Unless the appellate court believes it highly probable that the error did not affect the judgment, it should reverse.”
The College Hospital “reasonable chance, more than an abstract possibility” test is somewhat comparable to Traynor’s “high probability” test, yet superior, because it removes appellate courts from the guessing game of weighing abstract probabilities. Human beings, however, are accustomed to evaluating “reasonable chances.” We know the difference between a real chance and an “abstract possibility.” While we bemoan our fate when real chances are taken away from us, we usually do not give the matter a second thought when we lose out on an abstract possibility. Thus, a “reasonable chance” test is more concrete than a “reasonable probability” test.
Oliver Wendell Holmes reminds us that rules that can be concretely understood are better than abstract rules:
Does not a man remember a concrete instance more vividly than a general principle? And is not a principle more exactly and intimately grasped as the unexpressed major premise of the half-dozen examples which mark its extent and its limits than it can be in any abstract form of words?
Roger Traynor strongly believed that the Watson “reasonable probability” test was both too nebulous and too error-affirming. His classic essay concluded with an appeal for a new standard: “Unless the appellate court believes it highly probable that the error did not affect the judgment, it should reverse.”
The College Hospital “reasonable chance, more than an abstract possibility” test is somewhat comparable to Traynor’s “high probability” test, yet superior, because it removes appellate courts from the guessing game of weighing abstract probabilities. Human beings, however, are accustomed to evaluating “reasonable chances.” We know the difference between a real chance and an “abstract possibility.” While we bemoan our fate when real chances are taken away from us, we usually do not give the matter a second thought when we lose out on an abstract possibility. Thus, a “reasonable chance” test is more concrete than a “reasonable probability” test.
Oliver Wendell Holmes reminds us that rules that can be concretely understood are better than abstract rules:
Does not a man remember a concrete instance more vividly than a general principle? And is not a principle more exactly and intimately grasped as the unexpressed major premise of the half-dozen examples which mark its extent and its limits than it can be in any abstract form of words?
Thus, the Supreme Court of California and the California Courts of Appeal must now begin to regularly apply the “reasonable chance, more than an abstract possibility” test, in both criminal and civil cases, and while doing so, evaluate and discuss the appellate record in detail, and thereby provide both the judiciary and appellate counsel with concrete examples of the extent and limits of the rule. At the same time, the “reasonable probability” test should be discarded as ambiguous and confusing.
The “reasonable chance, more than an abstract possibility” test is also closer to the entire rationale of California’s 1911 constitutional amendment, i.e., reversals based upon unimportant and immaterial errors and technicalities are a “miscarriage of justice.” Yet side by side with that rationale there exists a deep rooted human feeling that when the law entitles a litigant to prove something, the party should be allowed a reasonable chance to do so. Denial of that opportunity is prejudice.
The “reasonable chance, more than an abstract possibility” test is also closer to the entire rationale of California’s 1911 constitutional amendment, i.e., reversals based upon unimportant and immaterial errors and technicalities are a “miscarriage of justice.” Yet side by side with that rationale there exists a deep rooted human feeling that when the law entitles a litigant to prove something, the party should be allowed a reasonable chance to do so. Denial of that opportunity is prejudice.
Lawyer Login