http://www.sv.vt.edu/classes/MSE2094_NoteBook/96ClassProj/examples/triplpt.html
and, at a temperature a bit below 0 Celsius, going from a pressure of about 400 atmospheres (which I think it would roughly be, at a depth of about 4000m) to 1 could indeed going from liquid to ice - I didn't know the ice-liquid boundary curved leftward as you go up in pressure from the triple point (or I had forgotten - doing a web search now, I see arguments about whether the pressure under ice skates is what melts a small amount of ice to allow them to slide easily, which I had heard of, back in school).
And what they did was allow the water to expand up the hole, wait for it to freeze, and then take it, lessening the chances of contamination:
Now it seems that perseverance may have paid off. By withdrawing the drill last year the scientists were able to allow the high pressure water in the lake to expand up into the borehole where it then froze. A strategy designed to minimize the chances of contaminating the lake from above.
...
The first core of transparent lake ice, 2 meters long, was obtained on January 10 [2013] at a depth of 3,406 meters. Inside it was a vertical channel filled with white bubble-rich ice,
http://www.huffingtonpost.com/2013/01/14/lake-vostok-water-ice-sample-antarctica_n_2471782.html