Context : Aims : We report on a spectral study at radio frequencies of the giant radio halo in A 2142 ( z=0.0909 ) , which we performed to explore its nature and origin . The optical and X-ray properties of the cluster suggest that A 2142 is not a major merger and the presence of a giant radio halo is somewhat surprising . Methods : We performed deep radio observations of A 2142 with the Giant Metrewave Radio Telescope ( GMRT ) at 608 MHz , 322 MHz , and 234 MHz and with the Very Large Array ( VLA ) in the 1–2 GHz band . We obtained high-quality images at all frequencies in a wide range of resolutions , from the galaxy scale , i.e . \sim 5 ^ { \prime \prime } , up to \sim 60 ^ { \prime \prime } to image the diffuse cluster–scale emission . The radio halo is well detected at all frequencies and extends out to the most distant cold front in A 2142 , about 1 Mpc away from the cluster centre . We studied the spectral index in two regions : the central part of the halo , where the X–ray emission peaks and the two brightest dominant galaxies are located ; and a second region , known as the ridge ( in the direction of the most distant south–eastern cold front ) , selected to follow the bright part of the halo and X-ray emission . We complemented our deep observations with a preliminary LOw Frequency ARray ( LOFAR ) image at 118 MHz and with the re-analysis of archival VLA data at 1.4 GHz . Results : The two components of the radio halo show different observational properties . The central brightest part has higher surface brightess and a spectrum whose steepness is similar to those of the known radio halos , i.e . \alpha ^ { 1.78 ~ { } GHz } _ { 118 ~ { } MHz } = 1.33 \pm 0.08 . The ridge , which fades into the larger scale emission , is broader in size and has considerably lower surface brightess and a moderately steeper spectrum , i.e . \alpha ^ { 1.78 ~ { } GHz } _ { 118 ~ { } MHz } \sim 1.5 . We propose that the brightest part of the radio halo is powered by the central sloshing in A 2142 , in a process similar to what has been suggested for mini-halos , or by secondary electrons generated by hadronic collisions in the ICM . On the other hand , the steeper ridge may probe particle re-acceleration by turbulence generated either by stirring the gas and magnetic fields on a larger scale or by less energetic mechanisms , such as continuous infall of galaxy groups or an off-axis ( minor ) merger . Conclusions :