How many innocent people are you willing to throw in jail in order to convict a guilty person?
It's a classic problem of criminal law. It's the foundation of why we require proof of guilt beyond a reasonable doubt. "Better that 10 guilty people go free than that one innocent person be imprisoned."
But what's the right number to let go free? 10? 100? 1000? Presumably that number gets reflected in our internal (or external) definition of what "reasonable doubt" entails. If we're only willing to let 10 go free -- if we're unwilling to let 11 roam the streets -- then presumably "reasonable doubt" means something like "90% certain". Whereas if the relevant number is 1000 or so, then reasonable doubt means something more like being 99.9% sure the guy's guilty. Because if that's the number, we can be sure that only 1 time in a 1000 or so will we be incarcerating innocent people.
Like I said, that's an age-old debate. One I've deliberately oversimplified a bit. If only because it's nearly impossible to summarize a century of jurisprudence and thousands of law review articles into a couple of paragraphs.
I mention this not simply to highlight a recurring problem of criminal law. But also because it's directly at issue -- despite the fact the California Supreme Court never once mentions it -- in this case.
It's a drunk driving case. Prosecutions that happen all the time. Where, if only due to the law of large numbers, we've undoubtedly subjected a number of innocent people to criminal punishment. Nonetheless, we clearly want to get it right.
There's a breath test in this case. The question in this opinion is whether an expert gets to testify about the potential inaccuracies of the test. The guy blows a 0.095 -- and then a 0.086 -- on the machine, and the relevant statute says you're guilty if you're driving with over a .08. Is it admissible for the expert to say not that this particular machine is inaccurate (e.g., improperly calibrated or the like), but rather that by its very nature the output of that machine is inherently inaccurate?
There's good reason, by the way, to believe that that number may well be inaccurate. A fact that nearly every scientist or person knowledgeable in the field recognizes. What we're trying to get at with breath tests is the underlying blood alcohol content (BAC). Because we don't care at all, obviously, how much booze is in your breath. We instead care about how much is in your blood -- which means, in turn, in your head. As that's what makes you a danger.
Is breath content a decent proxy for blood content? Yes. Yes it is. That's Henry's Law. In general, when you have a permeable membrane -- like the aveolar sacs in your lungs -- that touches a liquid (e.g., blood), stuff seeps through. So the concentration of booze in your breath is roughly the concentration of booze in your blood.
The key word, of course, being "roughly".
'Cause we also know, from extensive scientific study, that "rough" is really quite "rough". Which matters a lot when there's a sharp dividing line -- e.g., .08 -- between guilt and innocence.
One reason for the "roughness" is what's called the "partition ratio". Simply put, with some people, more booze gets through the aveolar sacs to the lungs, than with other people. This ratio varies from person to person. Women are generally different than men. Lung capacity (and condition) also plays a role. There is, in short, a ton of variability.
None of which matters to the Legislature. The Legislature has by definition set the partition ratio at 2100 to 1. We're going to declare -- despite the fact that we know it's not true -- that the amount of alcohol in 2100 parts of breath by volume is equal to the amount of alcohol in 1 milliliter of blood.
When you do that, by definition, you know you're convicting "innocent" people -- people who do not, in fact, have a blood-alcohol concentration of .08. Indeed, we've done studies, and we know almost exactly how many innocent people we'll throw into jail as a result. To its credit, the Legislature has set the 2100:1 ratio that -- for most people -- understates the ratio of alcohol in the bloodstream. (The average ratio is actually around 2300:1.) But for some people -- scientific estimates range from 2.3% of the population to as low as 0.3% of the population -- the Legislature's partition ratio nonetheless still overstates their individual ratio. A person like this will be convicted of driving with a BAC of .08 (i.e., will blow a .08) even though their actual BAC is less than that.
We know that. But we're willing to convict them anyway. Even when the relevant machine is working perfectly fine.
The proposed expert testimony here discussed a related, but similar, issue. Something that scientists also uniformly know. You see, what we're trying to get at with breath tests is the alcohol content in your actual aveolar sacs. But guess what? We can't get there. If only because we don't want to (or can't) stick a tube down there and pull out the stuff directly. So we get it indirectly. By making you blow "deep breaths" into the machine. That last little bit of air comes largely from deep in your lungs; e.g., the deep sacs. So that's a rough proxy.
But there's that word again. "Rough".
Because guess what? We know for a fact that the air that comes out of your mouth isn't the same as the air in your sacs. It's contaminated by the usual stuff even laypeople might know. Mouthwash. Alcohol in your oral cavity you might have just imbibed. Barf. Stuff like what. Which is why we generally wait a while to have you breath into the machine. To try to minimize those contaminants.
But what scientists know is that there's stuff you can't minimize. In particular, mucus and other liquid stuff in your upper lungs. That stuff may contain a fair amount of accumulated alcohol. It's liquid, after all, and has been repeatedly exposed to prior alcohol. So even if the breath in your aveolar sac is at .07, after it goes over this mucus and other stuff in your upper respiratory system, it might pick up enough alcohol to be .09. Which is what the machine might read. So you get thrown in jail despite the fact that your BAC is lower than .08 (i.e., is at a point at which we allow people to permissibly drive).
That's what the expert here wants to say at trial. But the California Supreme Court unanimously holds that he can't. Because, it says, the statute inherently defines the offense as having a breath concentration at the relevant point. The fact that this may have an imperfect -- or even utterly no -- correlation to what we in fact care about (i.e., the alcohol content of your blood) is irrelevant.
There's a lot to be said for that conclusion as a linguistic matter. It's actually right.
But it also says a lot about the age-old debate about "reasonable doubt".
One way to feel okay about incarcerating innocent people is to futz with the underlying concept. To say that it's okay to convict sometime even if you're only 90% (or 99%) sure they're guilty. Yes, we may thereby put an innocent person in prison. But that standard would nonetheless stop 10 (or 100) people from going free, so it's worth it.
The other way to solve this problem, however, is to do what the Legislature and the California Supreme Court have done here. It's to say that we don't care. To define the offense in a way that makes someone guilty even if we know they may not in fact have the characteristics about which we actually care. So we may actually think that .08 alcohol blood content is the dividing line between when you can and cannot be allowed to drive. But we nonetheless won't define the offense that way. We'll instead define it as .08 in the person's breath. That way -- by definition -- no one's innocent. Which is what the California Supreme Court says here. Which is in turn why the expert's testimony is inadmissible to the per se offense at issue.
Which some people may perhaps be fine with. After all, they're driving after having imbibed alcohol, and we're not that psyched about that, right?
Though realize that we could do the exact same things for other crimes as well. Worried about not being able to convict murderers? Just define the offense differently. Say that you're hereafter guilty of "murder" whenever you've (1) actually committed a murder, (2) been found with a gun in your hand within 50 feet of a dead person who's been shot, and/or (3) have confessed to a murder. Boom. No "innocent" people have been wrongly convicted. By definition. Despite the fact that it's the very definition of the crime that's the really troubling part.
We don't do that for murder. Or virtually any other crime. For good reason.
But we do here.
It's a neat little twist on the longstanding debate. A way in which the Legislature "avoids" the problem of "innocent" people being incarcerated by defining the problem away.
Which works. Except for the necessity of the quotation marks around the relevant terms.
