Perhaps I learned something that I really should have thought about before.
Some storm cells passed directly overhead (the flashes). But the thunder was not heard for several seconds. Assuming a speed of sound of approximately 340 metres/second, the interval between the flash and the thunder puts the distance at 2,000 metres or more. Hence, it was likely to be cloud-to-cloud lightning.
Not just that. The thunder was heard not as a loud bang but more of a rumble. This makes sense, as horizontal lighting (horizontal being a term of convenience rather than strictly accurate) means that the sound generated by different parts of the flash reaches the listener at different times, resulting in a rumble rather than a crashing bang.
So might it be, in general terms, true to say that thunder heard as a bang is cloud-to-ground, but when heard as a rumble, more likely to be cloud-to-cloud?