The Meaning of Yellow

I recently ran a post about "building bridges" with people with whom you do not get along. The discussion it provoked was interesting, with several people showing up to effectively demonstrate why attempts at finding common ground and building bridges can fail. Yesterday, I had a similar experience when I commnted at The Blackboard. It might be my fault as I am told:

You might want to think about how hostile and petty your contributions sound from the get go.

I'll admit I can be hostile and petty at times. As such, I won't rule out the possibility that statement is accurate. People who care can judge for themselves how petty and hostile this comment is:

I happened to see this post and felt I needed to point something out. I get the analysis in this post is hypothetical, but I feel it is important to note Oregon doesnt require a car clear an intersection to avoid being given a ticket for running a red light. I dont know of any state which does. There may be some, but in the case of Oregon, you only run a red light if you enter the intersection after the light has already turned red.
I couldnt find Jarlstons analysis with a quick search so Im not sure what assumptions he used, but based on the formula this post provides for his analysis, it would appear he has made a non-trivial error. The W/v term should not be included in any analysis of yellow light timing in Oregon.
For what its worth, I tried searching for states where this scenario accurately describes the traffic laws. I couldnt find any. Someone else might be able to. If not, it would appear all red light cameras (in the United States, at least) are triggered only if a car has entered the intersection after a light turns red.
Incidentally, red light cameras generally have a small grace period between the light turning red and them starting to trigger. Ive found reports of them ranging from .1-.3 seconds. I dont know if thats worth including in ones analysis.

As well as the tone of the follow-up comment I posted 15 minutes later:

As a quick follow-up, I found this document providing detail on the analysis in question. I see I was mistaken to say that term shouldnt be included in timing of yellow lights. The reason Mat Jarlston includes that term is because it is relevant for safety purposes. You want vehicles in an intersection while on red to clear the intersection before the next green light comes on.
That just has nothing to do with running red lights. You only get a ticket for it if the light is red when you enter the intersection. If you cant safely stop at a yellow light, you shouldnt get a ticket for running a red light.

I don't think these comments exhibit a great deal of pettiness and hostility. Maybe readers will disagree. I'm not going to worry about that in this post. You see, I've become a bit fascinated by the subject matter discussed in those comments. Originally, I only spent a little time researching this topic. When I wrote those comments, I thought they'd just be taken as a minor point of interest that wouldn't go anywhere.

That wasn't the case. Because of how the discussion played out, I wound up spending quite a bit more time reading up on traffic lights. I even talked to a couple family members who are over for the holiday weekend about traffic lights in a casual discussion. What I came to realize is there is a lot of justified uncertainty and confusion about what traffic lights mean. Given that, today I'd like to discuss a simple question, "What does a yellow light mean?"

For some context, this all came up because a man named Mats Jarlstrom criticized the timing of traffic lights because he felt yellow lights were not displayed long enough. While doing so, he called himself an "engineer." This got him fined and threatened with further penalties on the ground he is not allowed to call himself an "engineer." It's rather strange. You can read more about the case here if you're interested.

The technical issue Jarlstrom raises is relatively simple. I found an interesting description of the problem which includes a helpful image:

Put simply, it is possible for a traffic signal to switch from green to yellow at a point where a car cannot safely stop yet also not reach (and/or clear) the intersection before the light turns red. The area in which a driver will find themselves in this situation is called the "dilemma zone. The post at The Blackboard explains this arises because:

Evidently, currently traffic lights are timed using an equation that results from THE PROBLEM OF THE AMBER SIGNAL LIGHT IN TRAFFIC FLOW by published Gazis, Herman, and Maradudin in 1959 which is available here. In that paper, the authors analyzed the situation where a car moving at constant speed either brakes and stops or they maintain their constant speed and move through the intersection at their initial speed

I've seen similar claims in a number of reports of this controversy. The idea is this 1959 analysis assumed drivers would drive at a constant speed, meaning the timing of traffic lights could be off for a person decelerating at they approach an intersection. What makes this interesting is the description of the "dilemma zone" I showed above is from that 1959 paper. The paper is quite interesting and includes discussion like:

Approaching an intersection at a speed lower than the speed limit is one facet of defensive driving. It is seen from the preceding discussion that this in itself is not always sufficient to obviate the dilemma-zone problem. Another facet of such defensive driving consists of the maneuver of coastinig toward the signal light with one's foot readied on the brake. The advantage, in this case, which comes from shortening the reaction time, is reflected in a deerease of the critical distance XC. The improvement, which is by no means an absolute cure, can be seen from the curves plotted in Fig. 8 for two values of 62 other than the observed average. Such defensive driving, however, should be used with discrimination and great caution when approaching intersections in a high-density traffic pattern since it may induce a rear-end collision-a prominent type of accident in traffic today.

It is not clear to me how a paper which discusses the effect of slowing down as one approaches an intersection (and later, speeding up), would be used to time lights in a manner which assumes drivers travel at a constant speed. If that has happened as claimed, then people have misused the analysis presented in the paper. The paper shows 60 years ago people were considering the very same issue Mats Jarlstrom is now discussing.

I'd like to provide two additional quotes from the paper. First:

However, we believe that it is the duty of the traffic engineers and the drafters of traffic ordinances to present the average, honest, driver with a solvable decision problem. As it stands now, a driver who is in the middle of an intersection when the red light comes on may not be a deliberate violator, but may be the victim of an improperly designed light cycle.

And second:

We believe that a correct resolution of this problem may be found in one of the following alternatives:
1. Design the amber-light phase according to some realistic criteria in order to guarantee that a driver can always be in a position to obey the law.
2. If the amber-light phases are to be kept short relative to criteria such as determined herein, it may be desirable to state the vehicle code in such a way as to make it compatible with the driver, car, road, and signal characteristics.
In either case it would be very advisable to educate both the driving public and the law-enforcing agencies as to the exact operational definition of the amber light. Needless to say, the fewer the variations of traffic
ordinanices in this respect, from one locality to another, the fewer the chances of confusion. We wish to re-emphasize our hope that a wellthought-out and operationally sound traffic and enforcement system, together with the healthy driver attitudes of a properly educated public, will promote safer and more efficient driving conditionis

The reason I offer these quotes is I agree with the authors of the paper, especially when they say fewer variations in traffic ordinances would lead to fewer opportunities for confusion and thus improved safety on the roads. At the same time, the authors refer to people violating traffic ordinances by being in the middle of an intersection when the light turns red. That is interesting to me because everywhere I've lived, the law was if you entered the intersection before the light turned red, you were not violating any law.

A source for insight on this incongruity can be found in a Minnesota court case titled State v. Kilmer. In this case, a driver Justin Kilmer was stopped by a police officer in Minnesota in 2006 after he entered an intersection while the light was a steady yellow (as opposed to flashing yellow). The officer found Kilmer was intoxicated and cited him for driving while under the influence. A district court upheld this conviction/ The Minnesota Court of Appeals overturned it because:

The statute on which the district court relied in concluding that Kilmer violated a traffic law when he entered the intersection on a yellow light is Minn. Stat. 169.06, subd. 5(a)(2)(i) (Supp. 2005), which provides that [v]ehicular traffic facing a circular yellow signal is thereby warned that the related green movement is being terminated or that a red indication will be exhibited immediately thereafter when vehicular traffic must not enter the intersection . . . .
Subdivision 5(a)(2)(i) is one of three directly relevant traffic-control signal laws, all found in subdivision 5. Subdivision 5(a)(1)(i) treats the issue of a green light, stating that [v]ehicular traffic facing a circular green signal may proceed straight through or turn right or left . . . . And subdivision 5(a)(3)(i) deals with a steady red light, providing that [v]ehicular traffic facing a circular red signal alone must stop at a clearly marked stop line, but if none, before entering the crosswalk on the near side of the intersection . . . .
In concluding that Kilmer violated a traffic law by entering the intersection on a yellow light, the district court erred in two respects. First, the court relied on a case that applied a yellow-light statute in effect in 1944, but not in effect at the time of Kilmers arrest. The court relied on Flitton v. Daleki, 216 Minn. 549, 551, 13 N.W.2d 477, 478 (1944), and quoted its holding that Minn. Stat. 169.06, subd. 5(b)(1), prohibits entering an intersection after the yellow or caution light is shown. The version of the yellow-light statute in effect when Flitton wasdecided contained this prohibitory language: Yellow alone, or Caution, when shown following the green or Go signal(1) Vehicular traffic facing the signal shall stop before entering the nearest crosswalk at the intersection, but vehicles within the intersection may be driven cautiously through the intersection. Minn. Stat. 169.06, subd. 5(b)(1) (1941).
Subdivision 5(b)(1) was amended during the 1947 legislative session, and codified in 1949, to delete the mandate that traffic shall not enter an intersection on a yellow light and to replace it as follows: Vehicular traffic facing the signal is thereby warned that the red or Stop signal will be exhibited immediately thereafter and such vehicular traffic shall not enter or be crossing the intersection when the red or Stop signal is exhibited. Minn. Stat. 169.06, subd. 5(b)(1) (1949).
The subdivision has been amended three more times since 1947. A 1963 amendment, which was codified in the 1965 Minnesota Statutes, contained warning language that [v]ehicular traffic facing a steady yellow signal is thereby warned that the related green movement is being terminated or that a red indication will be exhibited immediately thereafter when vehicular traffic shall not enter the intersection. Id. (1965). Unlike the 1949 version, the 1965 version did not prohibit crossing the intersection when the red or Stop signal is exhibited. Id. (1949). In 1969, an amendment retained the language of the 1963 amendment regarding the warning nature of a yellow light but added an exception for proceeding on a red light when a green arrow is exhibited. Id. (1969). Finally, in 2005, the legislature amended the subdivision again but retained all of the language of the 1969 version, except the word shall in the phrase traffic shall not enter the intersection on a red light was replaced by the word must. Minn. Stat. 169.06, subd. 5(a)(2)(i) (Supp. 2005).
Thus, the yellow-light statute in effect when Kilmer was arrested, and which is currently in effect, contains no prohibition against entering an intersection on a yellow light but rather discloses the meaning and the advisory-warning nature of that signal.

I know that's a lot of text, but I think it's interesting as it shows how this issue can have a storied past. In Minnesota in 1944, simply being in an intersection when the light turned red was a traffic violation. Over the years, the law was amended several times, with the most significant change being in 1963 where the law was changed from prohibiting being in an intersection when the light turned red to merely prohibiting entering the intersection when the light turned red.

Forty four years after the law was changed in this way, a police officer stopped Kilmer for behavior which was legal. When a Minnesota district court reviewed the case, they upheld this action based dupon a 1944 law, failing to note the law had been changed multiple times and failing to realize it ceased to prohibit this behavior over four decades prior.

If police officers and court judges can be unaware of what a yellow light means, what does that mean for the average citizen? Suppose a person living in Minnesota moves to Lousiana. Should they be aware Lousiana's current ordinance states:

Vehicular traffic facing a steady yellow signal alone is thereby warned that the related green signal is being terminated or that a red signal will be exhibited immediately thereafter and such vehicular traffic shall not enter or be crossing the intersection when the red signal is exhibited.

The Lousiana law matches the design of the Minnesota law which existed from 1944-1963. Should citizens be required to educate themselves on information like this? I don't know. I can see arguments going both ways. What I can't see is how a person living in Oklahoma who knows the state ordinance states:

2. Steady yellow indication:

a. vehicular traffic facing a steady circular yellow or yellow arrow signal is thereby warned that the related green movement is being terminated or that a red indication will be exhibited immediately thereafter,

Should be aware if they drive through the city of Tulsa (where I once lived), the city has an ordinance which states:

Vehicular traffic facing a steady yellow signal is thereby warned that the red "stop" signal shall be exhibited immediately thereafter and such vehicular traffic shall not be crossing the intersection or signalized location when such red signal is successively exhibited.

While Sand Spring, a suburb of Tulsa, has an ordinance which says:

"Steady Yellow Alone":
1. Vehicular traffic facing the signal is thereby warned that the red or "Stop" signal will be exhibited immediately thereafter, and such vehicular traffic shall not enter or be crossing the intersection or pass the signal when the red or "Stop" signal is exhibited;

Can people living in Oklahoma realistically be expected to know the state ordinances on traffic signals are different than city ordinances? There are no warning signs. Many towns don't even post their ordinances online.

Suppose a person living in Oklahoma routinely sees yellow light and chooses not to stop (when they could do so safely) because they know the state law says as long as they enter the intersection before the light turns red, they are safe. Now suppose they are driving through Tulsa and get stopped for this behavior which was legal just fifty miles away. Is that fair?

While you consider that, imagine if the person drives through Oklahoma City instead of Tulsa. There, the city ordinance is:

Yellow or amber alone, "caution."
a. the showing of such signal color following green shall constitute a warning that the "red" or "stop" signal will be exhibited immediately thereafter.
b. vehicles facing the signal shall stop before entering the near side crosswalk or at the limit line, if it is marked, unless the vehicle is so near the limit line when the "caution" signal first flashes that a stop cannot be made in safety, in which event vehicles may proceed cautiously through the intersection.

In Oklahoma City, it is legal to be crossing an intersection when the light turns red, but if you could have safely stopped when you saw the yellow light, then you get in trouble. That's three different sets of rules a citizen would have to keep in mind while driving in OKlahoma.

So what does a yellow light mean? Does it mean you have to stop when you see it if you can do so safely? Does it mean you are simply being warned that a red light is coming soon so you may want to stop to avoid entering an intersection on red? Does it mean you are being warned a red light is coming soon so you may want to stop to avoid being in the intersection on red?

The answer seems to depend on where you are. Not only does it change as you go from state to state, but it changes as you go from city to city. You could drive along a road and see two consequtive yellow lights in five miles which mean entirely different things. That seems strange.

I don't know if the confusion this sort of variance could be used as a legal defense in court. I don't know if in cases like this courts might be forgiving of people who follow the state ordinances yet fail to follow a city ordinance which is different. I don't even want to think about what might happen if I had to drive through another country.

What I do know is from here on, I am going to assume every yellow light is "restrictive," meaning I Must stop at it if I can do so safely. I am also going to assume that at any given light, I might find myself caught in a "dilemma zone" and get a traffic citation for a situation I could not have avoided.

I get why people worry about federal governments overreaching their authority, but things like this make me wish we had national standards on basic things like traffic laws (especially since so much traffic is interstate). Then again, I've also said things like:

I haven't really done any research for this post. It's late, and tomorrow's Easter (Happy Easter everybody!). I just wanted to ask a question that was bugging me. Today I tried an experiment where I drove 15 miles an hour over the speed limit for half an hour. Not only did I never get stopped, I got passed by multiple people. If we can all drive 15 over the speed limit with impunity, what's the point of having a speed limit?

To further test this issue, I decided to drive exactly the speed limit on the way home. In an hour drive, I got passed by three police cars. If three police officers in a row decided going the legal speed limit was not required, why do we have a speed limit?

How does it make sense to have a law everyone agrees shouldn't be followed? And if everyone agrees it shouldn't be followed, how does anyone ever get a speeding ticket? How does a police officer justify breaking a law then giving a person a ticket for breaking the exact same law?

In the past. Plus, I am still baffled Michigan's traffic ordinances say

(b) Except when overtaking and passing on the right is permitted, the driver of an overtaken vehicle shall give way to the right in favor of the overtaking vehicle on audible signal and shall not increase the speed of his or her vehicle until completely passed by the overtaking vehicle.

As I don't think expecting people to honk while passing a vehicle makes a lot of sense in modern times. Maybe I'm the the weird one. Even if I am though, there's still an important question.

Do you know what the traffic laws are where you live? If you see a yellow light, what should you do?


  1. Judging from recent observations, drivers in my state believe that the yellow means one must hurry through the light so that the next two drivers can get through when it turns red. The third guy has to stop. If you are at the intersection stopped at the red in your direction, you must wait several seconds after it turns green to let the red light runners get through. This obviously isn't the written law, but what usually happens in practice. I know of no one who was ever arrested for running a red light.

    I wouldn't call you hostile or petty. Just a bit too fastidious and tenacious.

  2. Gary, I hope you don't know anyone who has been arrested for running a red light. I mean, getting a ticket for running a red light would be one thing, but arrested?! ­čśÇ

    Kidding aside, I have seen places that are like that. It always shocks me. I don't ever want to drive in one.

    And for the record, I don't think I'm petty... most of the time. Sometimes I am though.

  3. Steven Mosher, as a moderation rule at this site, comments which only contain links to or quotes from other material are not allowed. If you wish to refer to other material, you are required to add some commentary to explain what you think the material indicates and/or why people should care about it.

    I've told you this before, so the next time you do it, I won't release your comment from moderation.

  4. Huh. Well, at least Lewandowsky isn't using his old approach of nonsensical extrapolations to create nonsensical correlations. I suspect the paper's conclusions are rubbish though. People keep publishing studies in which they claim to show they can change people's opinions/views on things, but they always rely on asking people their thoughts shortly after whatever experiment they perform. That doesn't show there's been any change in view. If you indicate to people one answer is correct, they are more likely to regurgitate that answer to you. That doesn't mean their thoughts have changed or that they'd give the new answer if you asked them a week later. (And often, the "Changes" aren't quantified in an absolute way, resulting in readers not realizing the "changes" are quite small and unimportant.)

    I doubt I'll look into many details of this paper. I think people who accept this sort of study will believe it no matter what others say while most of the rest are naturally skeptical. I'd consider looking at it if the data for it were made available, but I already have one data request pending with John Cook. I suspect he won't provide the data I request in that one, and pursuing it will take enough effort.

    Interestingly, the data I'm requesting is from Doran & Zimmerman (2009). Peter T. Doran was a co-author on Cook's Consensus on Consensus paper, a paper which cited "consensus estimates" for three groups surveyed for the 2009 study. There were over 20 different groups. In total, results showing "consensus estimates" were published for only 7% of the respondents. As far as I can tell, the results for the other 93% of the respondents have never been published. That's a rather glaring omission from Cook's 2016 paper since it claimed to show all available consensus estimates. It's difficult to imagine how he could justify showing results for three particular groups that make up only 7% of a data set when the person responsible for the data set was a co-author on the paper. It's not like the data was unavailable to the team who wrote the paper.

    It's probably a waste of my time to try to get this data or the full results for the survey (I suggested the authors could provide tabulated results instead of the data if they preferred). Global warming is a major threat to human civilization, consensus messaging is so important in getting people to act to combat it, so clearly, data showing the consensus should not be shared...?

  5. The 'complete neutralization' is what sticks out to me. I suspect they started with the inoculation idea, and this decided what their result would be.

  6. Complete inoculation would be impossible. Heck, vaccines can't even completely inoculate you from viruses. There's no way some simple, short blurb of text would be that effective.

    But really, if you don't check back to see what (if any) effect your test had on the subject days or weeks later, you can't possibly know you've changed anyone's mind. It's easy to get people to give you different answers without changing their opinion on anything.

    Heck, you can get a lot of people to say black is white or that a square is a circle via things like societal pressure. There are tons of experiments showing this. My favorite is where you have four people give an obviously wrong answer then have the test subject provide his own. Many will give the wrong answer just because everybody else did.

    Actually, no. My favorite is the experiment where you have a group of people stand backwards on an elevator. Many people who board the elevator will also face the back. They think it's the expected behavior, so that's the behavior they exhibit.

    That doesn't mean they'll ride all elevators backwards from then on though.

Leave a Reply

Your email address will not be published. Required fields are marked *