One tends to think that any sizable impact would break the object apart. In that case, some of the material might eventually merge again via gravitational attraction, but structures like craters would very probably be erased.
Since this has not occurred, the question remains--why not?
It's been suggested that the comet could have been larger and denser when the craters were formed, and so, withstood the impacts.
It's also been remarked that the craters could be due to the comet expelling gas directly sublimed from its ice, rather than impacts. I'm not aware of any essentially circular areas of gas discharge in previously observed comets. It's already well known that impacts can create circular craters.
A third suggestion is that comet was impacted by other, similarly constituted objects, which limited the impact damage, due to their low mass and density.
Since the density of comet 67P/Churyumov-Gerseminko is very low in comparison to other comets, it seems unlikely that it would just happen to be struck by another comet of similar composition, especially where relatively large, rare impacts are concerned.
It also seems unlikely that gas expulsions would take the form of circles, like craters, rather than random shapes, particularly where large features are concerned.
The irregular shape of the comet suggests the possibility that it is a fragment of a larger object. Comets do sometimes break apart when they are near the Sun, due to loss of ice from heating. It appears, from the dynamical history of 67P, that it has spent the vast majority of its life, since the formation of the solar system billions of years ago, at a distance of at least 4 astronomical units from the Sun. A comet is not expected to be notably affected by the heat of the Sun at that distance, to lose much of its substance, or to be likely to break apart.
Edited by bison, 11 August 2014 - 04:09 PM.