SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1605.09096 · 2016

Diachronic Word Embeddings Reveal Statistical Laws of Semantic Change

arXiv · PDF · Open in the Atlas

Code that ran

6 of 8 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 6 have no contract verdict; 0 failures of our instrument; 2 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/1605.09096.json · badge: /api/ran/1605.09096.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
williamleif/histwords community (classifier label)
classifier basis not recorded on the link
4 ran · 2 no recorded run
network-embeddings/temporal_embedding_matching community (classifier label)
classifier basis not recorded on the link
2 ran · 0 no recorded run
FunctionStatusWhere it lives
align_list_of_embs Ran
ran on a synthesized input
network-embeddings/temporal_embedding_matching/align_embs.py
code served (permissive licence) · get_code("00d99d08ac61f19a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create embedding dictionaries with common keys
# emb_list: list of embedding dicts, each mapping keys to embedding vectors
# emb_base: base embedding dict with same keys

np.random.seed(0)

# Create base embedding with 4-dimensional vectors
emb_base = {
    'word1': np.random.randn(4),
    'word2': np.random.randn(4),
    'word3': np.random.randn(4),
}

# Create first embedding to align
emb_1 = {
    'word1': np.random.randn(4),
    'word2': np.random.randn(4),
    'word3': np.random.randn(4),
    'word4': np.random.randn(4),  # extra key not in base
}

# Create second embedding to align
emb_2 = {
    'word1': np.random.randn(4),
    'word2': np.random.randn(4),
    'word3': np.random.randn(4),
    'word5': np.random.randn(4),  # different extra key
}

emb_list = [emb_1, emb_2]

ARGS = {
    'emb_list': emb_list,
    'emb_base': emb_base,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
align_two_embs Ran
ran on a synthesized input
network-embeddings/temporal_embedding_matching/align_embs.py
code served (permissive licence) · get_code("8c86cccc3f85c086")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create small embedding dictionaries
# emb_to_align: dict mapping keys to embedding vectors
emb_to_align = {
    'word1': np.array([1.0, 2.0, 3.0, 4.0]),
    'word2': np.array([5.0, 6.0, 7.0, 8.0]),
    'word3': np.array([9.0, 10.0, 11.0, 12.0]),
}

# emb_base: dict mapping keys to embedding vectors
# Some keys overlap with emb_to_align, some don't
emb_base = {
    'word1': np.array([1.5, 2.5, 3.5, 4.5]),
    'word2': np.array([5.5, 6.5, 7.5, 8.5]),
    'word4': np.array([13.0, 14.0, 15.0, 16.0]),
}

# common_keys: explicitly specify which keys to use for alignment
# If None, the function will compute intersection automatically
common_keys = ['word1', 'word2']

ARGS = {
    'emb_to_align': emb_to_align,
    'emb_base': emb_base,
    'common_keys': common_keys,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
clean_pos Ran
ran on a synthesized input
williamleif/histwords/coha/cohastringutils.py
code served (permissive licence) · get_code("a01a035ff8870c96")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

def clean_pos(pos):
    tags = []
    for tag in pos.split('_'):
        tag = tag.strip('@')
        tag = tag.strip('%')
        tags.append(tag)
    return tags

ARGS = {
    'pos': 'noun_@verb%_adjective_@%adverb'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
clean_word_permissive Ran
ran on a synthesized input
williamleif/histwords/coha/cohastringutils.py
code served (permissive licence) · get_code("897b037a7b524b3c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re
import string

NUM = re.compile('^\\d*[-\\./,]*\\d+$')

PUNCT = set(string.punctuation)

AFFIX = set(["n't", "'s", "'d", "'t"])

def clean_word_permissive(word):
    if word == '@' or word == '<p>':
        return None
    elif word in PUNCT:
        return None
    elif word in AFFIX:
        return None
    else:
        word = word.strip().strip('*').lower()
        if NUM.match(word):
            word = '<NUM>'
    return word


ARGS = {
    'word': 'hello'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
marginal_r2 Ran
ran on a synthesized input
williamleif/histwords/statutils/mixedmodels.py
code served (permissive licence) · get_code("29c76d6ec4da6cc4")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

class StandaloneModel:
    def __init__(self):
        self.endog = np.array([1.0, 2.0, 3.0, 4.0, 5.0])
    
    def predict(self, params):
        return np.array([1.1, 2.1, 2.9, 4.2, 4.8])

class StandaloneResult:
    def __init__(self):
        self.model = StandaloneModel()
        self.fe_params = np.array([0.5, 1.5])
        self.fittedvalues = np.array([1.05, 2.05, 2.95, 4.15, 4.85])

res = StandaloneResult()

ARGS = {
    "res": res
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
process_lemma_line Ran
ran on a synthesized input
williamleif/histwords/coha/cohastringutils.py
code served (permissive licence) · get_code("41d3c7b00eb544c6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re
import string

NUM = re.compile('^\\d*[-\\./,]*\\d+$')

PUNCT = set(string.punctuation)

AFFIX = set(["n't", "'s", "'d", "'t"])

def clean_word_permissive(word):
    if word == '@' or word == '<p>':
        return None
    elif word in PUNCT:
        return None
    elif word in AFFIX:
        return None
    else:
        word = word.strip().strip('*').lower()
        if NUM.match(word):
            word = '<NUM>'
    return word

def clean_pos(pos):
    tags = []
    for tag in pos.split('_'):
        tag = tag.strip('@')
        tag = tag.strip('%')
        tags.append(tag)
    return tags

def process_lemma_line(line):
    info = line.split()
    if len(info) != 3:
        return (None, None, None, None)
    word = clean_word_permissive(info[0])
    if word == None:
        return (None, None, None, None)
    lemma = info[1]
    pos_tags = clean_pos(info[2])
    if pos_tags[0] == '':
        lemma_pos = None
    else:
        lemma_pos = lemma + '_' + pos_tags[0][:2]
    return (word, lemma, lemma_pos, pos_tags)

ARGS = {
    "line": "running run VB_VBG"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
make_data_frame Not yet run
unverified (harvested, no recorded run)
williamleif/histwords/statutils/mixedmodels.py
code served (permissive licence) · get_code("185e114a33b53568")
How it ran

no invocation recorded.

run_lmm Not yet run
unverified (harvested, no recorded run)
williamleif/histwords/statutils/mixedmodels.py
code served (permissive licence) · get_code("4f9bed08d41f5491")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 8 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Understanding how words change their meanings over time is key to models of language and cultural evolution, but historical data on meaning is scarce, making theories hard to develop and test. Word embeddings show promise as a diachronic tool, but have not been carefully evaluated. We develop a robust methodology for quantifying semantic change by evaluating word embeddings (PPMI, SVD, word2vec) against known historical changes. We then use this methodology to reveal statistical laws of semantic evolution. Using six historical corpora spanning four languages and two centuries, we propose two quantitative laws of semantic change: (i) the law of conformity---the rate of semantic change scales with an inverse power-law of word frequency; (ii) the law of innovation---independent of frequency, words that are more polysemous have higher rates of semantic change.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("1605.09096")
get_code_for_paper("1605.09096")
have("1605.09096")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/1605.09096.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/1605.09096.svg)](https://syntology.ai/paper/1605.09096)

Connect an agent — have() is free.