This question is really still up in the air.
I think much of this comes from the problems of tracking motion in three dimensions when the basis of motion is also moving. If you trace a line of where my hand moves relative to my trunk when I swim; you would find it is in an S shape from overhead to my hips. however remember that we rotate our trunks when we swim too. So you have to consider where one’s hand moves relative to the water. In this case you would find that my hands moves mostly straight back from catch to finish. Mostly, not totally.
Years ago doc counsilman got the idea that swimmers make an S during the freestyle pull. He did some studying and got the idea that this was because of the fluid mechanics involved. He applied bernoulli’s law of fluid motion to the system and got the idea that we create lift by moving our arms sideways relative to the water. It should be noted that as is the case in many swim technique studies, he was chasing the technique. He was trying to explain scientifically something that he saw good swimmers do.
Also remember that there weren’t the motion capturing devices we have now. It was just some cameras, some trigonometry and a grad student looking at pictures. So anyway doc counsilman came up with the idea that Bernoulli’s law applied and got people going on the lift theory. That we use our hands like the wings of an airplane and create lift by moving them sideways relative to the water.
There was a lot of research that followed; some arguments but the result was that the S curve came to be accepted and persists in a lot of swim circles. Mostly perhaps because once an idea gets into coaching and into red cross teaching; it stays there for a while. I think it’s a sociological issue actually. Kids get taught to make the s curve when they are age group swimmers. That’s what they learn when they are young and impressiionable and that’s they do. They get older. Older folks are less ready to change and already know plenty about swim technique so they head to the pool and get a workout in. While working out they are still pretty good so the guy in the next lane asks the 45 year old gentleman in the next lane what to do with his hands when he pulls so he says. Make an S in the water. The matter is of course confused by the fact that he may be correct.
Back to the story, much research goes into swimming by lift with lots of papers written and some guys doing research that refutes it but the community MOSTLY goes with the propulsion by lift idea. Lots of s curves. But perhaps it was the early nineties or the late eighties that fluid mecahnics became concerned with quasi steady state systems and shed vortices. The most notable being the bumblebee. Remember when we were told that according to the laws of physics a bumblebee’s wings were too small for it to fly. Well that’s not true anymore, turns out a bumblebee CAN fly (!); and we know how he does it too. Yet another case of the scientist chasing the technique.
Taking this new theory of fluid mechanics I tkink it was Dave Costill who popularized the idea that these quasi steady state systems and shed vortices apply to swimming. He published these ideas of propulsion, which still lean heavily on the ideas of doc counsilman, in his books. You see these books on the shelf at the local bookstore.
Round the late nineties some other research got popular though. That’s the research by guys who have been contradicting swimming through lift for a long time. And a couple of new ways of looking at things. One was putting little wool tufts on one’s arm anchored at one end. These act like little windsocks and indicate which way the water was moving. Turns out that through a large portion of the stroke, water moves down your arm, that is from the elbow to the fingers. This doesn’t go along with any other ideas out there. Another was the idea of using computational fluid dynamics to study the flow. In these studies they found no indication of either lift propulsion or shed vortices.
What does all that mean? In my opinion I think it is still up in the air. I think the physics of swim propulsion are still missing a key element, or we are looking at it in exactly the wrong way. The latest book by Ernie Maglischo, Swimming fastest (it’s “The lore of running” for swimming) makes the case that we move in diagonal directions to move water backward simply because it is the most effective way to do it, yet the lift and quasi steady state laws don’t apply. Simply newton’s second law. To an outsider this might seem semantic, “So I should move my hand sideways as long as I realize that Bernoulli’s law doesn’t apply.” But really it dictates the extent to which moving your hand sideways to the water is beneficial. It would seem not much sideways motion is appropriate - this would jive with the previous poster’s observation that some good swimmers bring their arms pretty much straight back.
But then again we are chasing the technique and there is also a chicken and egg problem. Good swimmers a re a product of their coaches. If the coach has told these gentlemen to bring their hands straight back then that is very likely what they do. Is it possible that we could go to another workout in another pool and find that the swimmers there are just as good but hve lots of sideways motion in their strokes? Almost certainly is that is what the coach tells them to do.
And what should you do - for most triathletes I say don’t worry about it. Triathletes have a lot of other ways to get better in the water usually that don’t involve taking sides in a fluid mechanics discussion. If you have a nice catch under water and get a good surface area to pull with you’ll get better whether you move a lot side to side or not. Furthermore we haven’t even talked about biomechanics; that could be the most important idea of all. That is, maybe good swimmers swim the way they do not because it the best way to grab the water but rather it is the best way to use the muscles to achiieve the movement.
Work on your catch and everything else should work out fine.
If you MUST work on your pull, get a copy “Championship swim training”. It has a whole chapter on that they author’s call sculling. traditionally sculling meant the purely side to side aspects of the pull. These guys include straight back pull aspects under the term sculling. If you practice their drills you should get your own idea of what works for you in terms of the shape of your pull.