Soru

Read the following academic passage concerning historical linguistics and computational methodologies:

[Paragraph 1]
For over a century, the comparative method has served as the empirical cornerstone of historical linguistics. By systematically cataloging regular phonological correspondences across basic lexical items in attested daughter tongues, practitioners reconstruct unrecorded proto-languages and establish definitive genealogical relationships among linguistic families. Yet this classical paradigm possesses an acknowledged horizon: because phonological erosion and lexical replacement steadily degrade recognizable cognates over time, the comparative method encounters an impenetrable temporal barrier—the so-called "time depth limit"—typically estimated at eight to ten thousand years before the present. Beyond this threshold, genuine ancestral signals become statistically indistinguishable from background noise, leaving the deeper evolutionary origins of human speech inaccessible to traditional reconstructive phonology.

[Paragraph 2]
In response to this intrinsic limitation, a cohort of researchers has championed Bayesian phylolinguistics, an algorithmic framework adapted from computational evolutionary biology. Rather than relying solely on manual qualitative reconstructions of individual sound laws, Bayesian methods model the gain and loss of cognate vocabulary sets across languages as stochastic, tree-generating branching processes. By applying Markov chain Monte Carlo algorithms to binary matrices of lexical traits, these models infer rooted language family phylogenies and generate probabilistic divergence dates. Enthusiasts argue that by incorporating explicit models of evolutionary rates across thousands of cognate sets simultaneously, Bayesian phylolinguistics can quantify historical uncertainties and potentially probe genealogical connections that predate the conventional temporal threshold of orthodox reconstruction.

[Paragraph 3]
However, prominent historical linguists maintain that computational phylolinguistics suffers from a foundational vulnerability: its structural inability to reliably disentangle vertical linguistic inheritance from horizontal borrowing. While biological organisms primarily transmit genetic data along discrete ancestral lineages, languages routinely exchange vocabulary, morphological features, and syntactic patterns through geographical proximity, intensive multilingualism, and socio-political contact. When Bayesian tree models process shared lexical items that actually arose through contact-induced convergence, they systematically interpret such traits as evidence of shared common ancestry, thereby generating spurious phylogenetic trees and severely distorted divergence dates. Consequently, detractors contend that without the rigorous phonetic filtration provided by the comparative method, purely statistical lexical clustering risks reifying historical artifacts.

[Paragraph 4]
This methodological tension should not be resolved through mutual exclusion, but through functional recalibration. Bayesian phylolinguistics cannot legitimately substitute for the comparative method; sound-law analysis remains irreplaceable for validating true cognacy and establishing genetic relatedness. Instead, the true primary value of computational phylogenetics lies in its capacity to serve as an objective hypothesis-testing apparatus within linguistic groups whose genealogical validity is already established by classical criteria. When applied to well-defined families, Bayesian modeling allows researchers to systematically evaluate competing spatial diffusion trajectories and test correlations with archaeological and genetic data, transforming historical linguistics into a disciplined partner in interdisciplinary prehistory rather than an autonomous chronometer.

Based on the passage, match each numbered paragraph to the statement that most accurately describes its specific functional contribution to the development of the author's primary purpose.