The more I lent money to poor women, the more I realized that credit given to women brought change faster than when given to men. But how would we find women borrowers in a country where most poor women had never even seen a bank and where 85 percent could not read our billboards or advertisements? We had to devise a series of tricks and techniques....

Because of the rules of purdah (literally "the curtain, the veil") we never dared enter women's houses in the villages we visited. So when I went to meet village women, I never knocked on their doors. Instead, I would stand in a clearing between several houses, so that everyone could see me and observe my behavior. I usually brought one of my female students with me. This go-between would enter the house, introduce me, and speak on my behalf. She would then bring back any questions the women might have. I would answer the questions, and back the girl would go. Sometimes she would shuttle back and forth for over an hour. We wasted a lot of time. Often the young girl would not catch all my ideas or the women's questions would get jumbled. It was an inefficient system.

One day, as I sat in a clearing between the houses of the Jobra village, the skies clouded and it began to rain. As this was the monsoon, the rain soon turned into a heavy downpour. The women in the house sent out an umbrella so I could cover myself. As the rain increased, one of the elder women took pity on the poor go-between. "Let the professor take shelter next door. There is no one there," she said. "That way, the girl won't get wet."

The house was a typical rural Bengali hut--a tiny room, without any electricity, a dirt floor, and no chair or table. I sat on the bed in the dark, alone, and waited. There were wonderful food smells I recognized: balam, the tastiest long-grain rice in Bengal, and kachu shak, a sort of long asparagus spiced with bay leaves, ground cumin, and turmeric. A bamboo wall and cabinets divided this house from the one next door, and I could hear the muffled voices of the women talking to my go-between. Every once in a while, the girl would come around the bamboo wall to tell me what the women had said. As she relayed their questions, the women would crowd against the bamboo divide to hear my answers.

After twenty minutes of this--hearing each other's voices, but communicating indirectly through the go-between--the women started bypassing my assistant and shouting questions or comments over the wall in their Chittagonian dialect. As my eyes grew accustomed to the darkness, I could make out human shapes staring at me through the cracks in the partition.

Then suddenly the pressure on the partition grew too great and part of it collapsed. Before they knew it, the women, once so shy, were sitting in the room listening and talking directly to me. Some hid their faces behind a veil, others giggled and were too timid to look directly at me. But we had no more need for someone to repeat all our words.

That was the first time I spoke with a group of Jobra women inside their house.