A homosapien asks,
“I’ve read one of your recommended books, Exercises for Common Hip and Shoulder Dysfunctions, awhile back. Dr Osar says that when doing rows, shrugs, bench presses, and pushups, he doesn’t recommend pulling the shoulders down and back, and rather keep them in neutral position, with slight protraction and retraction during the movement. What’s your opinion on this? I’m trying to explain to my peers that it is better off to do movements like Osar recommends alongside working on overhead work more frequently, but I don’t have good enough evidence to explain things.”
This was in response to my article Overhead Shrugs vs Regular Shrugs.
There are a lot of ways you could answer this. I’m just going to riff. Hopefully some part of it is helpful.
I basically agree with Osar there.
I’m even more lax with those movements though. When doing seated rows for instance, I don’t even tell people to think about their shoulders. I cue them to keep their torso perpendicular to the floor. “Don’t swing back and forth.” “We want the arms to move, but the torso to stay still. We’re rowing, but not like we’re in a row boat.”
Now, if a person is rowing and their shoulders are riding up into their neck (rhomboid dominant seated row), then I will cue them to not do that. However, that can be very different than cueing someone to pull their shoulders down. I’m trying to get the person to avoid a movement, not generate a movement.
-> If you look at the line of pull of the rhomboids, you can see how they pull in and up:
So, when a person is doing a seated row, and their shoulders come together and up, we know the rhomboids are part of that.
(As I hit on in the overhead shrugs article, of course, a person, like a bodybuilding oriented one, may purposely want to target that area. That’s a different discussion. We’re talking general shoulder health and keeping muscles balanced here.)
This is a very subtle difference in cueing, but makes big differences in how people do movements. When you tell someone to do a movement, they tend to do it with a very hard contraction. So, when you tell someone to pull their shoulders down and back, they tend to do it so much, those muscles get worked more than you probably want. Sure, you might want a person to work the lower traps more when they seated row to avoid that shrugging
-> Lower traps pull in and down:
but why trade shoulders-too-pulled-up for shoulders-too-pulled-down? When we’re aiming for neutral?
Because when someone has gotten good at pulling their shoulders down and back, they often do worse at the opposite…moving their shoulders forward and upward.
That’s truly the crux of it. Shoulder retraction, downward rotation and depression isn’t something people usually have trouble with, but shoulder protraction, elevation and upward rotation i.e. overhead motion is much more commonly problematic.
Why focus on practicing retraction, downward rotation and depression then?
But ramble-on-we-must.
–
In a great deal of shoulder issues, the person is having trouble with-
1) Overhead motion
and or
2) Elbow behind their shoulder motion (like the bottom of a bench press or push-up)
We’ve basically gone over 1) already. If you want to improve moving your arms, you don’t want to be practicing pulling your arms down!
-> It’s tautological, right? Try to raise your arms overhead while keeping your shoulders down and back. The cueing is oxymoronic.
-> Even in a seated row, while you may work the rhomboids less by pulling the shoulders down, you’re working a whole host of other muscles which…pull the shoulders down.
With 2), where the elbows move behind the shoulders, pulling the shoulders down and back can help in that movement.
For example, when you bench press, if your shoulder blade is anteriorly tilted, then by pulling it down and back, you will give your elbow more room to descend before the humerus goes into extension, which can make the anterior shoulder (where most people have issues at the bottom of a bench press) feel better.
Notice here how the person’s scapula is anteriorly tilted:
And because of that, the humerus is extended (elbow behind shoulder). And as the elbow moves behind the shoulder, the top of the shoulder presses forward, right where people often have pain:
It’s like a lever. Move one side of the lever one way, the other side eventually compensates by moving the opposite way.
We’re changing the mechanics of the long biceps head when we don’t need to.
Because while the short head connects to the scapula, and can thus be slackened or have no change in tension when the scapula anteriorly tilts,
-> Less appreciated about the biceps is its role in flexion of the humerus.
So, if the scapula anteriorly tilts, that would slacken the short head of the biceps i.e. its origin is brought closer to its attachment point.
However, if the humerus is brought into extension at the same time, that can lessen some of that slack.
(You also need to factor in what happens at the elbow!)
This isn’t really relevant. The point is it’s clear the short head isn’t, say, clearly mechanically changed from anterior scapular tilt.
But the long head, due to its origin site, doesn’t slacken like that. The origin runs laterally. You can imagine how that’s going to cause the tendon to compress the superior humerus in a different, unlikely desirable, manner. Think clicking, catching, etc:
However, just by pushing the bottom of the scapula forward i.e. getting it more into neutral / posterior tilt, the humerus is pushed forward:

Note how you can see all the fingers by the scapula now, because the inferior (bottom) scapula has been pushed forward, making them more visible.
The elbow is now more in front of the shoulder. Thus, it has more room to go into extension now (like the bottom of a bench press), without being behind the shoulder.
Think of it as the humerus has a neutral relative to the scapula. Scapula moves? Humerus compensates.
Look at the difference below. Let’s use a floor so it’s easier to see.
Bottom of a floor press; scapulae anteriorly tilted => elbow quite behind shoulder (humeri in extension):
Bottom of a floor press; scapulae posteriorly tilted relative to above => elbow much less behind shoulders (less humeral extension):
The range of motion is the same in the sense the hand is the same height above the torso, yet the range of motion at the shoulder joint is significantly different:
And a great way to irritate the shoulder is to load it when the elbow is significantly behind it. (Think how cops arrest someone by pulling the arms, and thus elbows, behind a person i.e. this movement is often used as a way to “disarm” someone.)
But notice what we did above. All we did was change how the scapula was tilted. We did not have to pull it back and down. In fact, we pushed it forward!
-> An easier way to think about the above may simply be we’re trying to make the shoulders less rounded. BUT, that means the top of the shoulders go back WHILE the bottom of the shoulders go forward:

Notice how you can see less of the letters on the person’s upper back as they become more upright. Because that part of their upper back is moving forward!
Just to hammer this home a little more. Your typical “computer posture”:
Notice when the person’s torso leans forward (left images), which coincides with scapular anterior tilt, how their elbows are now behind the shoulder ala the humeri are in extension.
-> This of course does not have to happen:
But leaning on your arms all day isn’t great for the shoulders either.
A big benefit people have found is when they think about “down and back” at the shoulders, they tend to also get some posterior tilt, which IS something many people could go for, and IS something that can help a e.g. bench press feel better and it IS a lot easier to cue “shoulders down and back” than “shoulders posteriorly tilted.”
But you do not need to squeeze your shoulders together, or depress your scapulae, to generate posterior tilt! (With that, just because you pull your shoulders down and back does not guarantee you’ll posteriorly tilt them, which is why some people have not found that cueing helpful at all.) And you get what you train. The issues that getting good at pulling your shoulders down and back cause offset the benefit of (possibly) getting better at posterior tilt.
-> A much easier way to work on the bench press movement is to instead do a floor press, so the person’s elbow can’t really move behind the shoulder to begin with.

Humerus does still go into some extension, but it is much less relative to bench pressing, and the floor is much harder than a bench press pad, making anterior tilt of the scapula much harder (a good thing). In a really sensitive shoulder, I’ll also place a pad under the elbow here, so the humerus can’t even reach neutral.
In comparison to,
If available, changing a person’s environment can do the cueing for you, and it’s a lot easier to change a movement when the person doesn’t even have to think about it.
Think about what a bench press is doing with the shoulders- flexing them (it brings the humeri forward). Nothing wrong with that.
But, why would we want to practice pulling our scapulae back and down when we are trying to flex our shoulder? Again, try to raise your arm in front of you (shoulder flexion) while also pulling the scapulae down and back. It won’t feel so great, because you’re not allowing the scapulae to reciprocate.
You can imagine if the scapula doesn’t start moving above, the top of the humerus simply jams into the top bones of the scapula, and everything in between those bones. In other words, purposely not allowing protraction and upward rotation to happen is purposely causing impingement.
Now, it’s true pulling the shoulders together can also impact the humeri in terms of flexion / extension. Notice when the shoulders are pulled together, the humeri move into extension:
That is, you can get the humeri further back (or down in the context of a bench press) without actually extending the humeri. Again, the scapulae can move the humeri.
But in my experience that benefit is offset by the fact you make flexing the humeri much, much more uncomfortable when you’ve pinned the scapulae into retraction. (Scapular stability, a phrase that can be quite alluring to many fitness professionals, is often not desirable.) And it’s hard to cue people something like “when your arms come down, pull your shoulders together. When your arms go up, protract your shoulder blades.”
Instead, I usually just don’t say anything and let people move their shoulder blades naturally.
-> Which is easier to do with DBs compared to a barbell, because the hands aren’t fixed. When protraction happens, the shoulder blades moves around the ribcage. The hand and arm can’t go along with this movement if they’re fixed.
Notice how the glenoid (where scapula and humerus meet) rotates anteriorly (and superiorly) as the humerus moves:
The humerus can’t reciprocate if it is fixed.
Notice how the olecranon fossa (pointy back part of person’s elbows) rotates as the person’s shoulder blades move:
(Your hands can come together when you DB bench press. They can’t when you barbell.)
–
The context this all happens in is critical. Doing some movements where the shoulders are being pulled together can be fine. But if it’s being cued over and over and over, and a person is never practicing e.g. upward rotation, it’s not surprising the person has gotten really good at NOT being able to move their shoulder complex overhead.
If a basketball player wants to improve their 3 point shot, they go practice it. But if they practice it so much they don’t also practice their 2 point shots, then we can’t be surprised if their 2 point shooting gets worse.
When we train people, we are always making them practice something. If we’re practicing them pulling their scapulae down and together way more than we practice them moving their scapulae forward and upward, then we can’t be surprised when people get bad at moving their scapulae forward and upward.
Viviana Vivienne Croccolo-Huwald
January 13, 2020
Please illustrate your recommendations regarding Neck Pain, Pain, Posture Assessment, Shoulder Pain, Theoretical Movement Science.
I think I have a dislocated right shoulder which I fractured with a fall circa five months ago and it has remained very painful.
Could you please post a video of the exercises which are recommended in order to remedy a dislocated shoulder.
Thank you for your time,
Viviana Vivienne Croccolo-Huwald
vivianahuwald121@outlook.com
b-reddy
January 13, 2020
Hey Viviana,
For one on one help, take a look at this: https://b-reddy.org/the-remote-client-process/
John
April 10, 2020
Hey Brian,
I read this right when you made it in January but I forgot to comment on it! Thanks for the clarification! Although, I’m still wondering how the shoulder blades should be during back squats and barbell bench pressing. A rehab clinician from Norway (Kjett Larsen) says that a lat spread with a neutral t spine is the most ideal for benching and maintaining scapular mechanics, but I haven’t seen a video done on it. He also says that the scapulae should wrap around the rib cage during a back squat, but there’s no video on that either. I know that your article says that if you do the movement a couple of times, it shouldn’t be a big deal as long as it’s not carried over to every other pattern, but I feel that it’s oxymoronic to purposely train with suboptimal movement patterns and trying to achieve healthy movement patterns.
Would love your opinion on this,
John
b-reddy
April 12, 2020
Hey John,
Those cues don’t resonate much with me. Trying to cue someone’s thoracic spine position while bench pressing is a reach, and I’m not sure how or why someone would try to get the shoulder blades to wrap around the rib cage during a back squat. Trying to get the scapulae to protract while holding the hands behind your back? Maybe I’m not interpreting what you’re saying correctly, but I don’t see value in that.
Overall you may be trying to think in a bit too rigid terms here.
If a person is struggling to achieve healthy movement patterns at the shoulder blade, which patterns are they struggling with?
If they’re struggling with protraction and or upward rotation, then a back squat -where the shoulder blades tend to be squeezed together to help make a shelf for the bar- you could say is not ideal.
If a person actually has trouble with their shoulder blade(s) being too protracted, too often (not a common issue but still happens), then you could argue a back squat may be helping train a healthy pattern i.e. getting some more practice with retraction.
Frankly, if you’re rather concerned with a person’s shoulder health, then you don’t have to think too hard to conclude bench pressing, I mean, why do it at all?
Odds are it’s the same thing for a back squat. Holding your hands that far back, on a sensitive shoulder, doesn’t make much sense. Whether the shoulder blades are in a good position or not.
Where you want to distinguish between ideal movement patterns for health and ideal movement patterns for an exercise. What’s ideal for lifting the most weight in bench pressing is not going to be friendly to the shoulder joint (or the lower back for that matter). You can also have ideal form in an exercise that will still never be ideal for the health of a joint. (It doesn’t matter how perfect a person’s deadlift form is, it’s still a lot of stress on the lower back.)
Because what’s healthy for the shoulders in regards to bench pressing is to not do bench pressing. To be clear, I’m not saying people should never bench press. I’m saying if shoulder health is the priority, then they shouldn’t be bench pressing. You bench press when a person is healthy and wants to have some fun. You don’t do it when a person is e.g. rehabbing a bum shoulder.
Lucas
May 18, 2020
Full range push-ups cause pain in my front shoulder, so I end up doing half push-ups instead, but still… I feel these really stress the front delts. How can I progress to full range push-ups pain free?
I see many people (especially on youtube) doing push-ups using handles or DBs to increase range of motion, going all the way down till their chest touches the ground, and they don’t get shoulder pain. What gives?
b-reddy
May 20, 2020
Some people weigh more or less than others.
Some people haven’t done enough wear and tear to their shoulders yet so they can get away with a lot more than an older person or one with a shoulder injury history.
Some people can get there but need to get their shoulders feeling good first, then gently progress their way into that.
Some people are naturally more flexible than others so they can get away with more ROM without pain.
Lot of possible explanations.
Here’s an example of a push-up progression I’ve used: https://b-reddy.org/a-program-for-push-ups-at-75-years-old/
apmega
September 25, 2020
Are there any good regressions for overhead shrugs for people that don’t have that amount of overhead mobility? Same question for pull aparts. Is it okay to do partial ROM pull aparts if you have a hard time horizontally abducting through full ROM?
I feel like I have low shoulders, my scapulae tipped forward slightly at rest, but I have a pretty flat thoracic spine. I am actively trying to find the right approach, as I often have a baseline of stiffness and soreness probably from some of these imbalances.
Thanks.
b-reddy
September 29, 2020
Overhead mobility progression: https://b-reddy.org/a-progression-to-lifting-your-arms-overhead-pain-free/
This is probably one of the only exercises I use that’s somewhat similar to pull aparts (but note there is no resistance with transverse extension here, the resistance is still only due to gravity, even if the person is holding DBs, which I’ll sometimes do): https://www.youtube.com/watch?v=ILwZXd5Kftw
I never have the person move into pain. If there is pain when the arms move back, we don’t go that far back. (That’s the progression: only go as far back as comfortable. Often what’s comfortable will improve with time.)
If a person is having issues getting their arms overhead, and they have pain with a pull apart motion, odds are I would not be doing any work which was pulling the shoulder blades together. As the progression article above hits on, if a person is having trouble getting their arms forward and upward, you have to ask why you would be doing any work which involves the person pulling their shoulders together and or downward.
Szyszak
September 27, 2020
Hello Brian,
I would like to make sure I understand correctly the topic u described above. Basically people should chase the scapula posterior tilt by working on the lower traps if they have anterior tilted of the scapula (elbow goes behind the shoulder – in this case subscapularis is stretched out and high risk on injury) but the movement shoulder goes naturally (cannot force it back and down?) And muscles which contributed to anterior tilting of the scapula is mainly pec minor, short head of the bicep and lats?
Great article btw, regards!
b-reddy
September 29, 2020
Be a bit careful trying to generalize that much. Keep in mind the context of the article. While I use the scenario of a person just changing their scapula position, and having the humerus follow that change, it’s not as if a person always has to follow that. For instance, if a person types a lot with hunched over (anteriorly tilted) shoulders, they may very well get into a pattern of shoulder flexion with anterior tilt (the man in a blue shirt typing photo in the post).
Ultimately, if you want to know if the scapulae are tilted, look at the scapulae, not the humeri.
Focusing on the lower traps isn’t something I ever really worry about. I instead worry about what movement I want. For example, I’ll sometimes see someone with rounded shoulders being told they need to work on their lower traps, but what if the reason for their shoulders being that way is simply their upper back is too kyphotic? They might be able to get their shoulders in neutral by just standing up straight from the upper back, using the upper spinal erectors.
Any muscle which connects to the front of the scapula and pulls it down is able to anteriorly tilt it.
Whenever a person is having trouble digesting this stuff, it can be very helpful to instead focus on movement, not muscles. I go into the muscles when writing and such because there is some value there, but if I’m with a client and trying to get them to move differently, odds are I’m not mentioning the names of any muscles.
Pat
February 28, 2021
Thank you, it really helped me with my not responding chest. Do you have any cues for other exercises?