The reason is infants are people and you gotta count them if you want to talk about actual life expectancy. Their lives may be short and pretty boring to study but it's just factually accurate, although it's better to add the context you just gave.

If your goal is to map and present life expectancy in any informative manner, than the inclusion of infant and early childhood mortality is straight up counterproductive. It practically becomes a graph on early childhood mortality over time, nothing else.