In recent years, the turbo principle has been widely applied to various decoding schemes. Here, we extend the general idea of iterative decoding to a communication scheme where arithmetic coding is used for source coding and a recursive systematic convolutional code is used for channel coding. The scope of this paper is to develop a low complexity SISO decoder for arithmetic codes using the SOVA algorithm principle and a trellis representation of arithmetic codes. The proposed approach enables iterative decoding and information exchange between the convolutional SISO decoder and the proposed arithmetic SISO decoder. Performance of the presented scheme are evaluated for an artificial source transmitting over an AWGN channel. Simulation results show that the proposed joint source channel iterative decoder leads to a significant performance gain in comparison with classical tandem decoding.