I can recall reading a book by OS Nock about the history of early colour light signalling. Apparently debates at the IRSE got quite heated between the two factions, those advocating for separate junction heads, and those advocating for junction indicators. The separate heads were preferred by traditionalists, as it followed traditional semaphore practice. However, it was claimed that this arrangement meant drivers were ignoring red aspects. In the end, it was the simplicity of junction indicators, particularly where signals had multiple routes, that won.
Early junction indicators had a vertical indication for the straight route. This was because they used strip lights, which were almost impossible to prove. The use of filament bulbs allowed the junction indicator to be proved, and so the straight ahead indication was dropped.
The Southern wasn't the only railway to use three lamp junction indicators, this was also the standard on the underground. The Southern also used miniature yellow aspects for junctions.
And of course, the ultimate in traditionalists, the GWR, used a totally different colour light signalling system, where the colour lights exactly replicated the night time indications of a semaphore stop arm above a distant arm!
And now we are going back to using double-headed signals for junctions. Albeit not for the junction signal itself, but for the signals in rear.
