Shoulder Alignment in Downward Dog: Is External Rotation the Best Cue?

Editor's note: The following article is intended to provide general information for yoga practitioners and teachers. It is not a replacement for the personal advice of a health professional.

When considering shoulder alignment in downward facing dog, what are the first cues that pop into your mind? If you’re like most yogis, one of these cues is likely to be “external rotation.” With occasional exceptions, the instruction for the upper arm bone (humerus) to rotate externally in the shoulder socket (glenoid fossa) in down dog is a foundational alignment rule taught in most yoga teacher training programs.

Because the anatomical action of external rotation can be cued without actually using those specific words, there’s a chance you have been instructed to externally rotate your shoulders without even realizing it. You may have heard the suggestion in down dog to “wrap your triceps back,” “turn your armpits to face each other,” or “spin the eyes of your elbows forward.” These are all cues intended to create external rotation of the shoulder joint—and there are many more!

Why Externally Rotate?

Why is external rotation such a foundational alignment instruction in down dog? The main reason relates to shoulder safety. When we rotate the humerus externally as the arm lifts overhead (in what is called shoulder flexion), we create more space between the soft tissues of the shoulder joint and the “roof-like” bony structure of the shoulder blade (scapula) called the acromion process.

Creating this space in the shoulder joint means that as the arm lifts overhead, some specific soft tissues (namely the rotator cuff tendons and the subacromial bursa) will be less compressed between the head of the humerus and the acromion process. [Ref]

The compression of the shoulder joint’s soft tissues as the arm flexes is known as “shoulder impingement.” When shoulder pain is present during flexion, this is often diagnosed by a medical professional as “shoulder impingement syndrome” (also referred to as SIS).

So in theory, the more space we can create in the shoulder joint as the arm moves, the less impinged our tissues will be. Therefore, with an intention to lessen or prevent shoulder pain, our yoga community has learned to emphasize the anatomical action of external rotation in not only downward facing dog, but also in all other asanas that involve shoulder flexion such as urdhva hastasana (arms overhead), virabhadrasana I (warrior I), and adho mukha vrksasana (handstand).

New Insights From Research About Shoulder Impingement

The shoulder impingement model has been used for decades in medical and health fields, and it has clearly influenced the yoga and movement world as well. However, insights from newer research have in recent years thrown the SIS model into question, with the result that many clinicians are now changing their perspective on shoulder pain.

The first major insight that gives us reason to be skeptical of the SIS model is that pain and tissue damage often do not correlate on a 1:1 basis. [Ref], [Ref], [Ref] This research is actually not new, but it has only recently become more widely accepted in the health and fitness fields. Countless studies using imaging techniques such as MRIs reveal that many, many pain-free individuals actually have tissue damage in their bodies—such as rotator cuff tears in their shoulders, meniscus tears in their knees, and disc herniations in their spines. [Ref], [Ref], [Ref], [Ref], [Ref], [Ref], [Ref] And by contrast, many people who experience pain in their body have no tissue damage inside at all. [Ref], [Ref]

This tells us that pain is a much more complex experience than we have traditionally understood. There are countless other factors besides tissue damage that can contribute to pain perception in the body. Because pain is extremely subjective and unique for every person who experiences it, there is no way to determine if a yoga student’s shoulder pain is actually the result of “shoulder impingement.” (And as yoga teachers, we are not qualified to diagnose pain anyway.)

For further reading on this topic, you may be interested in the introduction to pain science article that I wrote for Yoga International a few years ago.

A second reason to question the shoulder impingement model is the unconfirmed assumption that the contact made between the tissues of the shoulder and the acromion process causes tissue damage in the first place. [Ref] Shoulder tissues are in fact always impinged whenever we take our arm overhead. It’s a normal, unavoidable occurrence that naturally happens when the arms lift—whether we externally rotate our arms or not! [Ref] [Ref] In fact, soft tissues make contact with bone all over the body when we move. So, why does this particular spot receive extra concern and caution in the yoga world?

Additionally, a number of high-quality research studies suggest that surgeries intended to “fix” shoulder impingement syndrome by physically altering the acromion process to create more space in the shoulder joint are often no better at alleviating symptoms than simple exercise treatment programs designed to strengthen and mobilize the shoulder. [Ref], [Ref], [Ref], [Ref], [Ref] If SIS can be successfully treated without surgically altering the acromion process to make more space, then the cause of the pain must have been something other than tissues coming into contact with this structure.

One further point to consider is that our tendons are actually very strong, resilient tissues. Tendons are made of collagen fibers, and collagen has the strength of steel. [Ref] Collagen is usually only successfully torn or injured in instances such as high-force impact accidents or medical surgeries. By contrast, the forces involved in the shoulder flexion that we do in yoga are very low impact, and unlikely to be inherently damaging to our rotator cuff tendons.

This is all strong evidence suggesting that the popular model of shoulder impingement as a common cause of tissue damage and pain is flawed. (Just to be clear, this is not to say that people do not sometimes experience pain in their shoulder when they lift their arm; it’s simply suggesting that the cause of this pain is not necessarily soft tissue impingement.)

Down Dog Shoulders: A New Approach

As I mentioned earlier, most yoga teacher training programs have traditionally insisted on shoulder external rotation in downward dog because it is believed to minimize shoulder impingement. But as we have seen, “shoulder impingement” as a model for pain and dysfunction is quickly becoming outdated in the face of current research. It is perfectly normal for the soft tissues of the shoulder to come into contact with the bone above them as the arm lifts—this happens naturally in everyone’s shoulders as they lift their arms overhead, whether their shoulders are healthy or painful. And although pain associated with shoulder flexion can certainly happen, pain is actually a complex, multifactorial experience that is too often oversimplified in the yoga world. Besides the fact that there is little evidence to support shoulder impingement as inherently pathological, there are also a multitude of other factors beyond anatomy that can contribute to painful experiences, including psychological and social ones.

With this in mind, it is appropriate for us to re-examine our alignment rules for down dog. External rotation does not need to be emphasized for safety because the model on which this reasoning was based is not supported by current evidence. While external rotation is absolutely fine for the shoulders in down dog, there is also nothing inherently wrong with internally rotating the shoulders in this pose. In fact, everywhere along the spectrum between these two end ranges is an acceptable alignment for us to explore in our bodies.

So, if neither rotation is inherently good or bad, how do we determine which shoulder rotation to teach in down dog?

I would suggest that instead of making assumptions about the best positioning for our students’ shoulders, we should consider letting go of the need to cue shoulder rotation in this pose at all. Our students’ complex, sophisticated nervous systems understand how to self-organize their bodies much better than our yoga pose alignment rules do. What if we simply allow each student’s shoulders to position themselves in the most efficient way for that student’s unique body in that moment? Some students may prefer external rotation, but some may also feel better in internal rotation—and some will feel just right with no rotation at all.

Instead of being attached to the older model of external rotation as the only “safe” alignment to suggest to our student, we now have more options.

And in the case where a specific student experiences shoulder pain in downward dog, a productive strategy would be to help this student find a different way to bring their arms overhead that does not create pain. But instead of being attached to the older model of external rotation as the only “safe” alignment to suggest to our student, we now have more options. Perhaps bringing the student’s hands up on two blocks could eliminate their shoulder discomfort, or we might try experimenting with shoulder internal rotation, or even scapular elevation (lifting the shoulder blades up toward the ears). Because pain is complex, any number of changes to the pose could offer input that successfully changes the output of pain.

In general, I have found that the less I try to control and dictate my students’ shoulder positions in downward facing dog, the more easefully and efficiently they embody the pose on their own. Alignment is still a helpful tool, and I absolutely use it in many contexts. But the more I study modern movement science and learn about the complexity of the human body, the more confident I am in letting go of certainty around “correct” and “incorrect” alignment cues for many yoga asanas.