1: \begin{abstract}

2: In this paper, authors shall introduce a finite element method by

3: using a weakly defined gradient operator over discontinuous

4: functions with heterogeneous properties. The use of weak gradients

5: and their approximations results in a new concept called {\em

6: discrete weak gradients} which is expected to play important roles

7: in numerical methods for partial differential equations. This

8: article intends to provide a general framework for operating

9: differential operators on functions with heterogeneous properties.

10: As a demonstrative example, the discrete weak gradient operator is

11: employed as a building block to approximate the solution of a model

12: second order elliptic problem, in which the classical gradient

13: operator is replaced by the discrete weak gradient. The resulting

14: numerical approximation is called a weak Galerkin (WG) finite

15: element solution. It can be seen that the weak Galerkin method

16: allows the use of totally discontinuous functions in the finite

17: element procedure. For the second order elliptic problem, an optimal

18: order error estimate in both a discrete $H^1$ and $L^2$ norms are

19: established for the corresponding weak Galerkin finite element

20: solutions. A superconvergence is also observed for the weak Galerkin

21: approximation.

22: \end{abstract}