Let me answer you with the words of the known Pali scholar Wilhelm Geiger:
"A consensus of opinion regarding the home of the dialect on which Pali
is based has therefore not been achieved. Windisch therefore falls back
on the old tradition - and I am also inclined to do the same - according
to which Pali should be regarded as a form of Magadhi, the language in
which Buddha himself had preached. This language of the Buddha was
however surely no purely popular dialect, but a language of the higher
and cultured classes which had been brought into being already in
pre-Buddhistic times through the needs of intercommunication in India.
Such a lingua franca naturally contained elements of all the dialects,
but was surely free from the most obtrusive dialectical characteristics.
It was surely not altogether homogeneous. A man from the Magadha country
must have spoken it one way, and a man from the districts of Kosala and
Avanti in another, just as in Germany the high German of a cultured
person from Wurttemberg, Saxony or hamburg shows in each case peculiar
characteristic features. Now, as Buddha, although he was no Magadhan
himself, displayed his activities mainly in Magadha and the neighbouring
countries, the Magadhi dialect might have imprinted on his language its
own characteristic stamp. This language could have therefore been well
called Magadhi even if it avoided the grossest dialectical
peculiarities of this language. As Windisch has rightly pointed out,
after the death of the master, a new artificial language must have been
evolved out of the language of the Buddha. Attempts were made to retain
the teachings of the Buddha in authentic form, and to impose this form
also upon the portions which, although derived from the monastic
organisations of the various provinces, were gradually incorporated into
the canon. In connection with the designation of the canonical language
as Magadhi, Windisch also refers to the Arsha, the language of the
Jaina-suttas. It is called Ardha-Magadhi, i.e., "half-Magadhi." Now it
is surely significant that the Ardha-Magadhi differs from Magadhi proper
on similar points as Pali. For Ardha-Magadhi too does not change the 'r'
into 'l', and in the noun inflexion it shows the ending '-o' instead of
Magadhis '-e' at least in many metrical pieces. On the other hand, as I
believe to have myself observed, there are many remarkable analogies
precisely between Arsha and Pali in vocabulary and morphology. Pali
therefore might be regarded as a kind of Ardha-Magadhi. I am unable to
endorse the view, which has apparently gained much currency at present,
that the Pali canon is translated from some other dialect (according to
Luders, from old Ardha-Magadhi). The peculiarities of its language may
be fully explained on the hypothesis of (a) a gradual development and
integration of various elements from different parts of India, (b) a
long oral tradition extending over several centuries, and (c) the fact
that the texts were written down in a different country."
(from Wilhelm Geiger, "Pali Literature and Language", pp. 4-5)