Jason Crandell Vinyasa Yoga Method Twitter Jason Crandell Vinyasa Yoga Method Facebook Jason Crandell Vinyasa Yoga Method Instagram Jason Crandell Vinyasa Yoga Method Pinterest

Subscribe for a free arm balance guide
+ five free sequences!


A New Take on Twists

Marichyasana III

Spring has sprung and all throughout yogaland, studios are hosting “detox flow” workshops that emphasize deep twists and side bends. And, no doubt, the vast majority of these workshops will declare that you need to keep your pelvis “stable” when you twist in order to keep your lower back and sacroiliac joints safe. Although the word “stable” in this context is slightly misleading, teachers are trying to keep their students from turning their pelvis when they rotate their spine. This is the way I taught twists for over 15 years — adamantly no less.

This is the point in the narrative where I tell you that I’ve changed my tune. I no longer think that the pelvis needs to stay fixed when the spine rotates. In fact, I prefer to allow my pelvis to rotate a little bit in the direction that my spine is twisting. This means that when I’m twisting to the right, I allow my pelvis to rotate slightly to the right. I know, I know — I don’t like change either. And, keeping the pelvis fixed has been the conventional wisdom in yoga for a while. But, hear me out while I make a few points.

The first thing to consider is that your pelvis and spine work best when they work together. Technically, the spine starts on top of the pelvis. But, in reality the spine and pelvis are structurally merged via the sacrum and coccyx — not to mention all the ligaments, tendons, and other soft tissues that bind the spine and the pelvis together. Moving your spine and pelvis together in an integrated, cohesive way is one of the most effective ways to minimize injuries since cohesive motion distributes the mechanical stresses of yoga postures. Too much concentrated stress in the sacro-lumbar and sacroilliac regions is more likely to create injuries than stresses that are more evenly distributed. This means that if you don’t let your pelvis rotate in the direction that your spine is rotating, you are more likely to concentrate stress in the sacro-lumber and sacro-illiac region.

This doesn’t mean that you can’t move the spine independently. You can. It simply means that the spine and pelvis should be on the same page. The pelvis should be doing at least a little bit of what the spine is doing, especially in demanding postures.

The second point to consider: Look at the relationship between your pelvis and spine in your forward bends and backbends. Do you inhibit the pelvis from moving in these postures, or do you move the pelvis and spine together in these postures? Chances are you initiate your forward bends and backbends by rotating your pelvis and spine simultaneously. In fact, if you did forward bends and backbends and you kept your pelvis fixed, you wouldn’t go very far and you’d create an enormous amount of undesirable stress on your lumbar. It’s the same when you twist.

The last two points are simple and telling. First, when you allow your pelvis to rotate slightly when you twist, you’ll probably move and breathe more freely. Second, when I talk to my students and colleagues that are physical therapists and orthopedists, they agree that allowing the pelvis to rotate in the direction the spine is twisting makes more sense than not allowing the pelvis to rotate at all.

Now, again, I’m going to flip the script a little. I don’t want to say that rotating the pelvis with the spine is the new, singular gospel that every yogi should follow every time they practice. Rather, I want to dispel the notion that keeping your pelvis fixed while twisting is inherently safer, more stable, or more dynamic because it’s not.

Finally, I want acknowledge that after years and years and years of teaching postures one way, I changed my mind. All yoga teachers should afford themselves this right so that we continue to question our own assumptions. And, I want to give you a few postures to try so that YOU, yes, YOU, can determine the different ways that these two options feel in your body.

So, get your mat out, do a few hip openers to mobilize your hips and experiment with the following four twists. In the first phase of each posture, you’ll keep your pelvis fixed. In the second phase of each posture, you’ll allow your pelvis to rotate in the direction that you’re twisting. Take a few breaths in each posture and tune into the sensations of your entire body.


Parivrtta Utkatasana
Phase 1: Keep both knees in the same plane. This is the easiest way to assure that your hips are fixed (not rotating).

Phase 2: Allow the knee that you’re rotating away from to slide slightly forward. This allows your pelvis to turn into the twist slightly.


Prasarita Padottanasana
Phase 1: Place your hand on the back of your pelvis. Keep both hips level with each other. This will ensure that your pelvis is not rotating when your spine is rotating.

Phase 2: Keep your hand on the back of your pelvis and allow your pelvis to rotate in the direction that your spine is rotating. The hip on the side that you’re rotating toward will raise slightly. The hip that you’re rotating away from will slightly lower.


Revolved Triangle Pose
Phase 1: Place your hand on the back of your pelvis. Keep both hips level with each other. This will ensure that your pelvis is not rotating when your spine is rotating.

Phase 2: Keep your hand on the back of your pelvis and allow your pelvis to rotate in the direction that your spine is rotating. The hip on the side that you’re rotating toward will raise slightly. The hip that you’re rotating away from will slightly lower.


Marichyasana III
Marichyasana III
Phase 1: Sit with both hips equidistant to the front of your mat. Bend your right knee, draw your right heel toward your sitting bone and rotate to your spine to the right in marichyasana 3.

Phase 2: Begin with both of your hips equidistant to the front of your mat. Bend your right knee and draw your heel toward your sitting bone. Then, slide your left leg and left hip an inch or two further forward. This will rotate your pelvis slightly toward the right. Rotate your spine to the right to and do Marichyasana 3.


  1. Thank you for this refreshing look at the spine and pelvis relationship. In some of the provided photos, the depth of the twist appears to be greater when the pelvis is allowed to rotate with the spine. How do you approach/instruct exploring new depths to these poses?

    1. The depth of the twist in phase 2 is definitely greater. Fortunately, it doesn’t require additional instructions since that is the way students will naturally twist if they aren’t told otherwise. -Jason

  2. Thank you Jason. After years of practice I feel that twisting with the pelvis ‘fixed’ has been straining my SI joint and causing lower back instability. I’d love to hear your thoughts on updog to downdog. This transition of backbend to forward bend also seems a lot for the lumber/SI joint.

    1. This transition can be difficult for some bodies. The easiest way to manage it is to go from up dog to all fours. Engage your core, bring your pelvis and spine to a neutral position, then transition to DD. Hope this helps. -Jason

  3. Thank you Jason – I wonder if that for new students and especially students who spend the day sitting before coming to class that this is a safer option – I also consider that this approach fits with a vinyasa form of yoga to allow the body natural movement and flow – the images show a more fluid and comfortable configuration as opposed to leveraging against joints to force rotation Thank you to both you and Andrea for sharing X I believe that if I had been taught to allow the pelvis to move in the direction of the spine when I first began yoga whether I would have avoided years of S I joint torment ? I teach all classes this way now ☀️

    1. Great to hear from you, Anne. It’s tough to look backward on a practice and decide what would have made a difference if only you had done it. But, yeah, moving the pelvis and spine together in all postures is definitely a key step in keeping the region happy.

  4. Your article is a breath of fresh air for me. I am not a flexible person, so when I twist I need to twist my pelvis also a little bit and it makes me feel comfortable (I often do Phase 2 of the pose pictures posted above). However, my teacher totally disapproves and always tells me to always let my pelvis stay and it makes me feel a little bit uncomfortable. I will share this article on my facebook. Thanks for your point of view. Yoga should be enjoyed and make us feel comfortable.

  5. Commenting on Rachel’s comment on transition from updog to dd being tough transition for lumbar/si joint…what has worked for me is to neutralize my spine before going straight to dd by coming up or pushing up on a plank with top of my toes still firmly planted on the mat from updog, thereby neturalizing my spine before going to dd.

  6. I LOVE when a teacher revisits their cues and is not afraid to say they found a better way. This teaches our students we are human and we are seekers, hopefully allowing them the power to change their minds too.
    Thank you for sharing such great tips on this site!

  7. Wondering about the knees in phase 2 of parivrtta utkatasana? My knees give me pain in uttanasana or utkatasana when my knees edge too far forward; it’s much better for me to keep slightly more weight in the heels and keep knees in line with the toes. Would extending the knee outside this alignment cause instability here?
    Fabulous article, much appreciated insight!

    1. Hey Ruth,

      You’re 100% correct. For some students, flexing the knee more deeply in this pose can be problematic for the knee. If this motion is creates discomfort in your knees, you should avoid it. Feel free to keep your weight a little further back and maintain a more vertical alignment of your shins!

  8. Hi! I am a yoga teacher and doctor of physical therapy student in Vermont. A friend of mine shared your blog with me and asked me what I thought. I am not a biomechanist, but I have studied a lot of kinesiology, anatomy, and functional movement. Here are my thoughts:

    1. the pelvis does not rotate
    look at the photos, the pelvis (the illium, ischium, pubis and sacrum) moves as 1 unit. in the images shown with a “stable pelvis” the woman pictured is shown keeping her pelvis stable in relation to her legs (legs stay in 1 spot, hips stay in 1 spot, spine twists alone)… in the images where the pelvis is “rotating with the spine” the pelvis is moving on the femurs (you can see this in the parivrtta prasarita padottanasana pose, the left side of the pelvis has moved anteriorly, inferiorly and rotated inwardly on the left femur while the right side of the pelvis moves superiorly, posteriorly and has rotated externally on the femur…. something simillar is also happening in revolved triangle and marichyasana, except in marichyasana the left femur has just abducted more while the right had adducted more) as the spine is twisting…. and/ or the joints of the lower legs have just moved more in one leg than the other (you can see this in the utkatasana example, the left knee is just flexing more than right, and the left ankle is dorsiflexing more than the right)….

    I think this is a place a lot of yoga teachers get confused, thinking they are seeing movement at one joint (like maybe the SI Joint) when they are really seeing movement at another (the hip joint).

    2. what is the real difference in the shear forces present at the SI Joint in one version vs another?
    I don’t know the answer to this question… BUT one general rule I have learned is:

    3. when the legs are in an asymetrical position (like, in a lunge, in a pigeon pose, or in these twisting poses the way he is suggesting they be done here) the potential for shear forces at the SIJ increases. this is why true SIJ injury only really occurs in people who do a lot of very assymetrical things with their hips (like, people who do yoga) or people with very lax ligaments (pregnant ladies for example, or people who have stretched the crap out of their ligaments doing things like… yoga with poor core stabilization)….. most “SIJ pain” in the general population is pain reffered from the lumbar spine

    this fact leads me to believe that Jason is probably proposing a method of twisting that will produce higher shear forces at the SIJ, not lower shear forces (though again, I can’t say this for 100% sure)

    4. Here is why I like to keep the hips symmetrical in yoga poses (twisting poses, lunges, pigeons, etc):
    It requires really good core stabilization to do so.
    That’s it in a nutshell. To me, having to hold the hips neautral while moving the trunk into a twist requires a lot of great core awareness, strength, and coordination, and those are all really good things to foster

    5. After studying biomechanics and functional movement here is how my take on twists in yoga has changed:

    I am currently a firm believer in NOT using anything but core strength to enter a twist. A lot of times in yoga I remember being encouraged to use the strength of my arms to drag my spine into a deeper rotation. I do not think this is a good idea (particularly because intense rotational forces can be damaging to the intervertebral disks).
    Here is how I teach twists now – When the legs are fixed and the spine is twisting (which is the case in all the poses Jason has pictured above), hold the legs/hips in a symmetrical shape and, WITHOUT USING YOUR ARMS, twist. then set your hands down wherever they have landed. You will end up in a shallower twist (safer and more functional) and you will have used a ton of core strength to get there (good!).
    When the spine is fixed and the legs are moving (for example in supine twists like windshield wiper ) all of the forces involved are less so I think the stakes are lower and the “how to” of the poses matters quite a bit less … as long as (again) your aren’t using gravity plus your own strength to wrench yourself deeper into a spinal rotation.

    Those are my thoughts! Although would not teach the twisting method Jason is suggesting here, I LOVE the ongoing conversation about “alignment” and biomechanics and modern postural yoga, and I think we all benefit from questioning what we have been taught. So, thanks for that Jason!

    1. Hey Margaret,

      Thanks for reading the article and sharing your expertise. I really appreciate it. And, honestly, I wish we were in the same room, talking, and looking at people in twists. But, alas.. this will have to do. I have some follow-up thoughts regarding your intelligent comments.

      1) Yes, the pelvis works as one unit. Agreed. Except, of course, for the small–but problematic–flares and tilts that can occur in the ilium.

      The Coxal joint rotates. As you wrote–and I described–the pelvis moves on the femurs. I’m calling the movement of the pelvis in which you turn the pelvis in the direction that the spine is moving “rotation.” The pelvis rotates in the transverse plane. It can move clockwise and counterclockwise. In fact, when describing the movement of the pelvis in your note–which you took even further than I did and included an expert description of femoral motion–you described the pelvis as rotated. So, I’m genuinely not sure why you wrote that the “Pelvis doesn’t rotate.” It’s possible that you meant the Sacrum doesn’t rotate. But, this article doesn’t contend that the sacrum rotates. It only expresses the concern that turning the spine without turning the pelvis (even a little) may disrupt and destabilize the connective tissues of the region, promote greater laxity in the region, and produce less congruity between the sacrum and the spine.

      3) Here here. I agree with your concern about asymmetrical postures with regards to SIJ issues. I’ll take this further and say that it’s not just asymmetrical postures, it’s the transitions in and out of asymmetrical poses that can be problematic for a lot of students.

      But, being concerned about asymmetrical postures and twists as they relate to SIJ issues aren’t mutually exclusive. In fact, I’m including a link from an article by Judith Laster, whose a Dr. of Physical Therapy and Yoga Teacher with are than 30 years experience in the field, that takes the same perspective as I do regarding the pelvis and spine in twists. http://www.yogajournal.com/article/yoga-101/anatomy-101-understanding-sacroiliac-joint/. In the article, she makes the exact same argument about initiating pelvic rotation prior to spinal rotation in twists. In fact, she made the argument LONG before I stated the case in this blog piece. And, in her experience, she believes that “twists are a top culprit behind SI joint injuries.” She too believes that allowing the pelvis and the spine minimizes SI risk.

      I’m never going to disagree with someone saying that they believe in less torque and more moderate movements in yoga. I’m 1,000 percent in agreement that a more “shallow” degree of rotation–and focusing on core stabilization–is a good idea. Yes, absolutely.

      4) I agree. In my experience, core stabilization and pelvic rotation aren’t mutually exclusive. I think students can be taught to do both.

      5) Again, this is totally reasonable. And, it’s largely what I advocate for in the first phase of twisting. Once the core is integrated, my preference, though, is to free up the motion of the pelvis and spine a little bit more than you prefer. I don’t crank away, but I do like a little more motion. Facilitating strength and stability is super important. Facilitating integrated mobility is also super important.

      Thanks again for your time, honesty, and well-reasoned thoughts. Your students are in good hands.


      1. Hi Jason,

        Thank you so much for your thoughtful and thorough response, it would be an awesome conversation to have in person I agree! After reading your thoughts and also reading the Judith Laster article (I am familiar with her work and respect her opinion as very well educated), I had a few more thoughts of my own, here they are:

        1. NOW I get what you mean when you say the “pelvis rotates”, you mean the pelvis rotates on the femurs, I was thinking you were referring to rotation at the SIJ (I have read things from other yoga teachers who think there is a lot more movement at the SIJ than there is – lots of researchers contend that there is none, those who think there is some agree that it is small enough to be unmeasurable by any tool currently available to us). Sorry, my misinterpretation of your words there!

        3. I would be curious what the evidence is that twists are a “top culprit behind SIJ injuries”, is this just Judith’s anecdotal experience, she sees that people get a lot of pain in that region in twists? I don’t see any references listed with her article. I know there is great evidence to show that rotational movement in the spine is a top culprit in disc degeneration and wear and tear on the spinal ligaments, but I have never heard this put forth as a mechanism of injury or degeneration at the SIJ. There is moderately good evidence to suggest that asymmetrical movement is a culprit, and I agree with you that this definitely includes transitions in and out of asymmetrical poses (by evidence I mean randomized controlled trials or that explore joint movement and joint stress).

        4. When I say “core” I am referring to the muscles of the gluteal region as well all of those of the abdominal region and the lumbar region (all of the muscles that control femoral rotation as well as a number of trunk rotators) …. I know the word “core” is not very well defined but that is what I meant by it.

        My main point of contention with the method you suggest is that I have never seen a student that had any difficulty rotating their pelvis on their femurs during rotational movement (twisting) of the spine, I have seen a lot of students (almost all of mine) for whom it is difficult or impossible to not turn their pelvis when they rotate their spine. Controlling rotational movement of the pelvis using core musculature while performing spinal rotation is an advanced spinal stabilization exercise we use often in PT, and that’s how I like to use my twists in yoga too (I usually teach twists as repetitive movement not as held poses). In my view, the only reason to teach a student to rotate the pelvis on the femurs in a twist is to allow them to get deeper into a twist without stressing their spine out as much (which I think is exactly what you and Judith both suggest). What I am suggesting is throwing out the idea of twisting that deeply in the first place because it serves no functional purpose (45 degrees of thoracolumbar rotation is considered healthy by the AMA, and more is seen as a potential source of pathology – of course that number is not exact but an estimate based on how individuals without spinal pathology move – in yoga classes it is common for up to DOUBLE that amount of rotation to be encouraged or even required for some “deeper” twists… I use the quotes because “deeper” seems to often imply “better” in yoga and I am not at all on board with that, and I think it is awesome that you are not either)

        I actually think in a lot of ways you and Judith and I are on the same page here (we all want to find a way for students to rotate less in their spines), I’m just saying; why solve the problem by kicking the rotational stress down the kinematic chain? Why not just quit hooking our arms around the back of our thighs all the darn time? Why find ways to shimmy ourselves into shapes that encourage dysfunctionally excessive movement? What’s the point? Unless someone wants to win at a yoga competition, or be a contortionist, or impress their friends I don’t see that there is one.

        Thank you for engaging in this conversation!

        1. Thanks again for your follow-up, Margaret. I’m going to keep this one brief and move all the way down to your last 3 paragraphs. And, we both agree that we’re largely on the same page.

          With regards to “core” I have a fairly extensive guide to yoga and the core on my site. I’m certain that we think about the core in very similar ways.

          With regards to spinal stabilization and pelvic rotation over the femurs, again, I don’t think these are mutually exclusive. I agree that students don’t have difficulty rotating the pelvis over the femurs, but the problem–as I see it–is that this normal, natural, functional rotation is coached out of students. When we tell our students NOT to rotate the pelvis over the femurs, I think we’re asking them to do something that is less natural and less functional. My preference, then, is to encourage the natural movement AND encourage core stabilization with twists. Again, I don’t see these as mutually exclusive.

          With regard to arm-positioning and postural intensifications, again, my preference is to facilitate stability and mobility. I’ll never discourage someone from producing less motion and more stabilization in a posture. And, I totally respect that you and other teachers prefer a lower-range of motion with decreased application of leverage from the upper-extremities. That’s a sound preference. But, to be frank, I also like some assisted motion and the whole-body proprioceptive experience of connecting my arm to my leg in twists.

          Thanks again for all your thoughts and your contribution to this practice!

  9. Hey Jason! Great article, and welcome to the “let the pelvis follow the lumbar spine during twists” team! I’ve been teaching revolved triangle and other twists this way for years, and have never really understood the argument for keeping the pelvis level. I’ve been meaning to write or create a video to illustrate the benefits of letting the pelvis rotate in twists for awhile now, as I share it with student in my anatomy courses, and your blog inspired me to finally put something out there. Check it out when you have a moment and let me know your thoughts 😉 http://www.asfyt.com/blog/alignment-of-the-pelvis-in-revolved-triangle

    1. Hey Jason,

      Nice to hear from you. And, thanks for your feedback. I’m happy it inspired you to create such a good video!

  10. I learned this concepts years ago from Judith Hanson Lasater and have been suggesting to students for years to explore the option of the lower spine region (Sacral area) to rotate with the spine..

  11. I’d really like help on how to make my body more open so I can go deeper in the actual hip postures e.g. to get my bent front leg straigther in pigeon,or when it’s split dog to get that bent leg further over.


  12. Jason,
    I’m wondering if you could offer your thoughts on how teachers can offer adjustments better to students in seated twists when allowing some rotation of the pelvis. Do you feel it’s sometimes (or always?) injurious to help a student ground through the sitting bones in a twist because that prevents the pelvis from moving in the way you are suggesting? How do you adjust seated twists these days?
    Thanks for your thoughts!

Add Comment