According to Fortunatov's 1895 theory, the verbosity in the "Proto-Lithuanian-Slavic" language shifts the stress from the preceding syllable if the articulation did not have an extension.
In Russian and Lithuanian the word bar̃zdą "beard" had no accent shift since the ending of the accusative case has an intermittent length.
[2] For the Lithuanian language, Vladimir Dybo introduced a clarification to this law: "The accent was shifted from the circumflex to the next final acute.
Criticizing Vermeer, Sergei Bolotov and Mikhail Oslon reject his explanation by means of a analogical restoration and try to derive a certain rule explaining the occurrence of the acute.
Thus, probably for the first time, Thomas Olander drew attention to the non-acute of the previous syllable: [I]t is perhaps more likely that the acute pre-LI *-āˀ- first became circumflex by analogy with the other stems and then was subject to Saussure's Law.
The appearance of the circumflex in place of the probable acuteness (from *-ah₂-) is due to Nikolaev's metatony, and not to the Olander's analogy.
Subsequently the Moscow Accentological School, after a thorough analysis of imaginary and marginal exceptions to the de Saussure's law, cancelled Dybo's clarification and introduced a reduction in the endings of primary cases or Leskien–Otrębski–Smoczyński's rule.