Thursday, December 9, 2010

When does an intervention begin? Story of Dr. Kiran Bedi’s first day at Tihar Jail

My engagement with my clients is called an intervention. It is similar to your doctor or fitness coach intervening in your regular routine through your working engagement with them. The objective is the same: help client become healthier. The question is: When does an intervention begin? There are two schools of thought. One school says that an engagement has two parts. The first one is about diagnosis which in an organization setting begins with surveys and/or interviews of various stakeholders and ends with planning of the intervention. The second part is the actual intervention. Prof. Edgar Schein belongs to the second school which says that intervention begins with the diagnosis itself i.e. with the first question you ask your client. Now, you might say, “What difference does it make?” Following story tells how much of a difference it can make.
The post of Inspector General (IG) of Tihar Jail, India’s largest prison was lying vacant for many months when Kiran Bedi was posted there. By then she had finished nine months of “paid wait” period after a full tenure as Deputy Inspector General of Police in Mizoram, in the North-East of India. The appointment was more like a “punishment posting”. One of the ex-IGs (Prison) told her, “What will you do there? There is no work there! I was IG (Prisons) many years ago; I received just two files a day. So I used to clear them from my home.”
Monday 3rd May 1993 was Kiran’s first day at Tihar. She had made a brief visit the previous Saturday and met her direct reports. Without settling down Kiran went for a round of Prison No. 1 (there are 4 prisons in Tihar). It was just a 20-yard walk from her office. She had to pass through two giant gates before actually entering the prison wards. Wearing a uniform was not mandatory. So Kiran wore a full-sleeved pastel pathan suite topped by a waist-length Nehru jacket. “This gave me a full cover, with a sense of grace” She writes. The Superintendent of the jail, K. R. Kishore, followed her. There was no armed guard. She held a notepad in her hand to record on-the-spot observations.
As Kiran filed past the waiting prisoners, the Warders, perhaps from the force of habit, started to physically contain the prisoners without the slightest provocation from them. Some even waved their sticks menacingly the onlooking prisoners, in a gesture to show concern for her security. Kiran signaled the Warders to stop doing this.
There were blank stares all around her. Kiran stood there not knowing what expression would be most suitable for the moment. Not being in uniform had already communicated a desire of informal tone. She started wondering if our system was at all designed to help change offenders and forgive those who were willing to mend. Perhaps in continuation of the thought, she suddenly broke the silence by asking them: “Do you pray?”
No one answered.
She repeated: “I am asking you, do you pray? Please tell me.” She spoke in Hindi.
The men looked towards the Warders as if to ask them if they were permitted to speak. The Warders were confused.
Kiran moved closer to the bunch and directed the question to one inmate chosen at random.
He answered, “Yes, sometimes,” nodding his head.
“Very good. Who else does? You?” She pointed to another prisoner, again at random, getting even closer to the crouching men.
And then one after another, voices joined in saying, “Yes, I also do. I recite the path. Most of us pray at our own timings…”
Perhaps the first human contact was made. She probed on, “Would it be better if ‘we’ say a prayer together? Would you like to?”
They fell silent again. They had never prayed together.
Then one of them, with one by one the staff and the other on me, said hesitantly, “Yes…” Others nodded their heads in agreement, wanting to be part of the prayer.
She said: “All right, which prayer should ‘we’ sing together? Can you suggest one?”
“Do you know ‘Aye Malik tere bande hum, aise hon hamare karam, neki par chalen (O Lord we are your creation/ May our actions be worthy)?” She asked.
This time there was an enthusiastic and instant response, “Yes!”
She said: “Get up to sing together.”
“Close your eyes and sing with me”
And they sang. Kiran says, “When our eyes opened, theirs and mine, I felt we had together succeeded in giving out the first signal of mutual trust which could set the pace and for our work relationship from now on”
Imagine if Kiran had sent out a survey through the Warders as the first step to gather the data!
Source: “It’s always possible: Transformation of one of the largest prisons in the world” by Kiran Bedi, Sterling publishers.


  1. Hi Vinay,

    1. your blog has captured the essence that intervention starts with the first contact whatever be the step.

    2. Some might do a survey / steps of diagnosis or some might (like in this case)do what is called " action research" where you act and use that data to diagnose and continue to act accordingly (more an evolving model of intervening).

    3. For all the interventions my opearting principle has been this statement from particle physics" The moment you observe it changes".


  2. How does an intervening person/agency generate confidence that the intervention is not thrust upon from outside, is beneficial mutually in a transparent way and will definitely be effective? This is perhaps the first step of any intervention that determines success or failure (even for the first step of doing the survey).

  3. Good question, Ravindra. Developing “Mutual trust” is an important first step. However, there should be a common ground between the two parties for trust to develop. For that to happen, each party should share its assumptions about the current situation and possibly the desired situation. However, it may not practical to do this in every situation. For example, if you take “teaching to a class of 50” as an intervention, it may be difficult in each class to understand “readiness of each student to receive the knowledge in that class”. In an extreme form intervention – one country going on war against another – helping may not be the objective. Motive may be to establish what one perceives as “justice” or “helping the people of ailing country”. No matter what, it is unlikely that there is a common ground established. In my consulting engagement, I try to do this.

    What is your experience?


  4. Hi Vinay,

    Actually, I was struck by the fact that in the incident that you narrated, how Kiran Bedi did everything together, creating a mutual trust, seeking information and the intervention itself. That's why, I think, the incident stood out. And you are right, this may not be possible always and one needs to systematically build the common ground even before the mutual trust. But I think, why Kiran Bedi could do it, the reason is more a humane appeal/angle on her part during the intervention, for example by her not donning the uniform and then creating trust with her informal conversation. What was the common ground and mutual trust between her and the inmates? Maybe the intervention itself when she started singing the prayer. Or was it something else? So, it takes us back to the question, where the intervention actually start and how?

    In a corporate set-up, take a person who has say at-least 5 years of work-exp, and when the intervening person/agency (external or internal) tries to seek input from such a person, the first question which springs in his mind is, "what are they going to do with this information and why do they want it now"? It's a loaded question, the doubts being raised on the motive (making people more protective in divulging) and doubts on effectiveness/utility of it all as well (making people indifferent), thus making survey only partially truthful. The proof of pudding is always in eating, so maybe the intervening person will always have to do a small act/presentation and prove/convince how it will be useful even before the survey starts. Dr. Kiran Bedi did it best, and I think, the same principle applies everywhere, whether it is in corporate world or an undergraduate class where a teacher should show some usefulness/application of the concept even before teaching that concept.

    This discussion and the setting of Kiran Bedi's incident reminds me of a beautiful movie by V. Shantaram "Do Aankhen Baarah Haath" and an urge to see it again. :-)

    thanks and regards,

  5. Ravindra,

    Really appreciate your inputs. It is helping me understand this concept better.

    As we see the initial communication between Kiran and the inmates wasn’t explicit. A lot got communicated through the dress-code and starting with a humble inquiry. In fact, Edgar Schein beautifully presents how an ideal helping relationship develops in his book “Helping”.

    According to Schein, when such a relationship begins there are two challenges: First, status of the helper and the one being helped are unbalanced. i.e. the former is starting at “one-up” and the latter at “one-down” positions. In Kiran’s case this gap is even wider as she is way up in the hierarchy. Second, chances are high the helper doesn’t know enough about the context of the one being helped. For all you know, the one being helped is just a cog in the wheel and unless you can find more about the cog and the wheel you can’t help meaningfully. In Kiran’s case, this is manifested in her wondering, “If our system was at all designed to help change offenders and forgive those who were willing to mend.”

    Schein suggests that in order to overcome the two challenges the first invention must always be what he calls “humble inquiry” even if the inquiry is merely careful observation and listening in the first few moments of encounter. Through such an inquiry, the helper is accessing her ignorance. It also makes a beginning in reducing the gap between the statuses between the two.

    While a survey is an inquiry, it is difficult to make it “humble”. As you mentioned, questions like, “what are they going to do with this information and why do they want it now?” is going to make the gap between helper and the one being helped only wider.

    Thanks again,