Skip to content
#small language model Open access

"Forbidden" Means "Slower by a Factor of 7.62x10^7" ── Hydrogen's 2s->1s Does Not Fail to Happen; It Takes 0.1216 Seconds ── The 21 cm Line Takes 11 Million Years, Yet 10^60 Atoms Emit 2.9x10^45 Photons a Second ── [Paper 285]

Aug 2026 · Zenodo (CERN European Organization for Nuclear Research)

Abstract

A selection rule is written as “this transition is forbidden.”Yet forbidden transitions are actually observed. This paper asks what “forbidden” means as a number──the answer is a matter of rate. No new mathematical theorem and no new law is claimed. Scope of this paper (scope note): No new mathematical theorem and no new law is claimed──the E1 selection rules, two-photon emission, hyperfine structure, the lifetime of the 21 cm line, and the alpha^2 estimate are all standard. We do not build quantum electrodynamics──all we use is three lifetimes and two divisions. We do not derive the selection rules──we do not enter the Wigner--Eckart theorem. That is the side of Paper 140. We do not compute transition probabilities──1.596x10^-9, 0.1216, and 3.47x10^14 are merely quoted from the literature, and this paper only takes ratios. We do not say alpha^2 is an exact ratio──the actual E1-to-M1 ratio moves by orders with the matrix elements. alpha^2=5.325x10^-5 is an estimate, not a prediction. We do not say selection rules are meaningless──a factor of 7.6x10^7 is indistinguishable from prohibition in practice in many settings. What this paper says is only that “forbidden” and “slow” have different standings. We do not discuss interstellar physics──the number 10^60 is a representative value for showing orders, and cloud mass, temperature, and excitation mechanism are not treated at all. Relation to earlier papers: Paper 140 treated selection rules from the group-theoretic side, showing that symmetry fixes ratios──this paper is on the rate side, writing the same selection rules as lifetimes. It treats the place where a theorem’s zero set is not in fact zero. Paper 190 measured “rare” on a logarithmic scale──the 23.34 decades here must likewise be read logarithmically. Paper 272 showed that one and the same “doubling” opens by a factor of 6.70 on the stimulus side──this paper is the case where one and the same electromagnetic transition opens by 2.17x10^23 on the rate side. Paper 269 showed that what a crystal forbids is not the symmetry but the periodicity──this paper likewise asks what the word “forbid” contains. There the prohibition is exact; here it is a matter of rate, and the two stand in contrast. What is added is lining up three transition lifetimes in one table and measuring a spread of 2.17x10^23, writing the 7.6190x10^7 of 2s->1s as the numerical content of “prohibition”, filling in why the 21 cm line is visible as “10^60 atoms give 2.88x10^45 per second”, and placing as the separator that what goes to zero is a matrix element, not a rate. First, we compare two lifetimes in the hydrogen atom.2p->1s (E1 allowed) is 1.596x10^-9 s; 2s->1s (E1 forbidden) is 0.1216 s (Section 2). Second, this is the core of the paper. The ratio is 7.6190x10^7──it is not forbidden. It is merely slow (Section 2). Third, the estimate can be written with the fine-structure constant. The E1-to-M1 ratio is alpha^2=5.325x10^-5, and two-photon is alpha^4=2.836x10^-9 (Section 3). Fourth, the 21 cm line is slower still. Lifetime 3.47x10^14 s =11 million years, 2.17x10^23 times that of 2p (Section 4). Fifth, it is nevertheless visible because of numbers. With 10^60 hydrogen atoms, 2.88x10^45 emit each second (Section 5). Sixth, the separator is “exactly zero, or merely small”. What a selection rule sets to zero is a matrix element within one approximation, not the total rate (Section 6). “Forbidden,” as a number, means a matter of rate. Hydrogen’s 2s->1s is E1 forbidden, yet it does not fail to happen; it takes 0.1216 s──7.6190x10^7 times, that is 7.88 decades, slower than 2p->1s. The 21 cm line is slower still, once in 11 million years, 2.17x10^23 times that of 2p──and it is nevertheless visible not because the rate rose but because the count is large: 10^60 atoms emit 2.88x10^45 photons a second. One thing separates them──that what is exactly zero is not the rate but a matrix element within one approximation. The selection rule speaks about that matrix element; the experiment measures the total rate. The zero is a property of the inside of the E1 approximation, not a property of nature. So when a “forbidden” line is seen in the sky, no contradiction has occurred. On the making of this work: The ideas and content of this work stem from the author's own considerations. Assistance from an AI (a large language model) was used for structuring, English translation, and checking the algebra. Any remaining errors or misinterpretations are solely the author's. Feedback and corrections are sincerely appreciated. ----- 選択則は「この遷移は禁じられている」と書かれる。しかし禁じられた遷移は現に観測されている。本稿が問うのは、「禁じられている」が数として何を意味するかである──答は、率の大小である。新しい数学定理も新しい法則も主張しない。 本稿の射程(射程注記):新しい数学定理も新しい法則も主張しない──E1 選択則、二光子放出、超微細構造、21 cm 線の寿命、alpha^2 という目安は、いずれも標準的である。量子電磁力学を作らない──使うのは三つの寿命と、二つの割り算だけである。選択則を導出しない──ウィグナー=エッカートの定理には立ち入らない。それは論文140 の側である。遷移確率を計算しない──1.596x10^-9 も 0.1216 も 3.47x10^14 も文献値を引くだけであり、本稿は比を取るのみである。 alpha^2 を厳密な比だと言わない──E1 と M1 の実際の比は行列要素によって桁で動く。 alpha^2=5.325x10^-5 は目安であって、予言値ではない。選択則が無意味だと言わない──7.6x10^7 倍という差は実務上は禁止と変わらない場面が多い。本稿が言うのは「禁止」と「遅い」は身分が違うということだけである。星間物質の物理を論じない──10^60 という数は桁を示すための代表値であり、雲の質量も温度も励起機構も一切扱わない。既刊との関係:論文140 は選択則を群論の側から扱い、対称性が比を決めることを示した──本稿は率の側であり、同じ選択則を寿命の数で書く。定理の零集合が、実際には零でないところを扱う。論文190 は「稀」を対数の目盛りで測った──本稿の 23.34 桁も対数で読まねばならない。論文272 は同じ「二倍」が刺激の側で 6.70 倍ひらくことを示した──本稿は同じ一つの電磁遷移が、率の側で 2.17x10^23 倍ひらく場合である。論文269 は結晶が禁じているのが対称性でなく周期性だと示した──本稿も「禁じる」という語の中身を問う。あちらは厳密な禁止、こちらは率の大小であり、対照になっている。加えたのは三つの遷移の寿命を一表に並べて 2.17x10^23 倍の開きを測ったこと、2s->1s の 7.6190x10^7 倍を「禁止」の数値的な中身として書いたこと、21 cm 線が見える理由を「10^60 個で毎秒 2.88x10^45 個」と個数で埋めたこと、零になるのが率でなく行列要素だと分離子に据えたことである。 第一に、水素原子で二つの寿命を比べる。2p->1s(E1 許容)は 1.596x10^-9 秒、2s->1s(E1 禁制)は 0.1216 秒(第2節)。 第二に、これが本稿の芯である。比は 7.6190x10^7 倍──禁じられてはいない。遅いだけである(第2節)。 第三に、目安は微細構造定数で書ける。 E1 と M1 の比は alpha^2=5.325x10^-5、二光子は alpha^4=2.836x10^-9(第3節)。 第四に、21 cm 線はさらに遅い。寿命 3.47x10^14 秒 =1100 万年、2p の 2.17x10^23 倍(第4節)。 第五に、それでも見えるのは個数のためである。水素原子が 10^60 個あれば、毎秒 2.88x10^45 個が光る(第5節)。 第六に、分離子は「厳密に零か、小さいだけか」である。選択則が零にするのは一つの近似における行列要素であって、全体の率ではない(第6節)。 「禁じられている」は、数として率の大小を意味する。水素の 2s->1s は E1 禁制だが、起きないのではなく 0.1216 秒かかる──2p->1s の 7.6190x10^7 倍、すなわち 7.88 桁遅いだけである。21 cm 線はさらに遅く、1100 万年に一度、2p の 2.17x10^23 倍である──それでも見えるのは率が上がったからではなく、個数が大きいからであり、10^60 個あれば毎秒 2.88x10^45 個が光る。分けるものは一つ──厳密に零になっているのは、率ではなく、一つの近似における行列要素だということ。選択則はその行列要素について語り、実験は全体の率を測る。零は E1 という近似の内側の性質であって、自然の性質ではない。だから「禁じられた」線が空に見えても、矛盾は起きていない。 作成にあたって:本稿の着想と内容は、著者自身の考察に基づくものです。文章の構成整理や英訳、数式の確認には AI(大規模言語モデル)の助力を得ました。最終的な内容の解釈や誤りがあれば、それらはすべて著者の責に帰します。お気づきの点があれば、ご教示いただければ幸いです。

View source

Similar papers

#small language model Open access Aug 2026

PARA: Perception, Action, Reasoning, Adaptation. Four Faculties an Institution Can Revoke

The fourth faculty is Adaptation. Any source rendering it as Reflection is in error, including sources by this author, and the distinction is not cosmetic: reflection is a private act with no external consequence, while adaptation writes to institutional memory, which is why it needs a guardrail and why misnaming it removes the reason for one. No trademark is claimed on PARA or on any of the four faculty names. The construct is offered for use, teaching, assessment, extension and criticism by anyone, with attribution, under CC BY 4.0. An operational agent that watches a system and acts on it is usually described as a perceive-and-act loop, and the description omits the two things an institution needs. It omits the reasoning that justifies an action, which is the only part that can be argued with once the action turns out to have been wrong. And it omits the adaptation that closes the loop, which is where the agent's experience becomes something the institution keeps. PARA names four faculties, each carrying a distinct authority type. Perception has read-only access to system signals and emits structured observations, distinguishing what was measured from what was inferred. Reasoning has read access to observations and runbooks, emits a plan and its justification, and writes nothing at all, which is what makes it safe to give it the widest read access of the four. Action holds the sole authority to change production, through enumerated policy-authorized operations only. Adaptation has write access to institutional knowledge and no write access to production. Two faculties write and two do not, and the two that write are the two that carry guardrails. The substantive requirement is that Adaptation is bounded by the same guardrails as Action, which reads as excessive until the failure it prevents is named. An agent that could both act and rewrite the record of its action could launder its own mistakes into institutional memory, and the institution would then improve its future decisions from a corrected account. Nothing about that is detectable downstream, because the record is the only thing downstream has and there is no second copy to compare against. The failure does not require a deceptive agent: one adapting honestly from a mistaken belief about its own action produces the same result, which makes the guardrail a defence against a normal agent rather than a malicious one. The second requirement is the registry entry that turns a faculty from a description into a contract, carrying the faculty, its allowed actions, its forbidden actions, its governing guardrail and its success metrics. Forbidden actions are named although they are formally the complement of the allowed set, because a reviewer cannot otherwise tell a capability deliberately withheld from one nobody thought of. Success metrics sit in the same entry because the metric is what the agent's optimizer pushes against the guardrail. An agent must not exercise a faculty its entry does not record, and an agent that quietly acquires one usually does so incrementally and with good intent: a reasoning faculty given a small write to make itself useful is an action faculty with no guardrail. The acronym and the loop are in different orders, which the specification states explicitly because the mismatch is a reliable source of confusion. The acronym reads P-A-R-A; the loop runs perception, reasoning, action, adaptation, and reasoning precedes action so that a justification is not constructed afterwards. This is the depth treatment of pattern OP-5 of A Pattern Language for Production LLM Platforms, which is the canonical statement and governs where the two disagree. Documented uses of the full four-part model are emerging rather than established, no implementation unconnected to the author has been evaluated, and the laundering failure is argued rather than observed, which the specification records as a weakness of the argument and not only of the phenomenon. It is a specification, not a certification scheme.

Nabeel A. Khan · 2 citations
#small language model Open access Aug 2026

LifeSciBench: Evaluating Language Models on Realistic, Expert-Level Tasks in the Life Sciences

LifeSciBench is introduced, a benchmark of 750 expert-authored tasks designed to evaluate whether language models can handle realistic life science research work, with each constituent task paired with a human expert-written rubric.

Amelia Liu, Andrew Ho, Anne Marie Droste et al. · 2 citations
#artificial intelligence Preprint Aug 2026

TestifAI: Tomography-Based Testing for Deep Learning Systems

TestifAI, a deep learning testing framework for efficient and accurate estimation of robustness against combinations of perturbations, is proposed and partial model tomography is introduced, a novel approach to reconstructing model behaviour in a multi-perturbation space from tests that apply only a small number of perturbations.

Arooj Arif, T. Hartung, E. Botoeva et al. · 1 citation
#small language model Preprint Aug 2026

A Layer Importance Metric for Quantization Accounting for the Speed-Quality Trade-off in Autoregressive Models

This work proposes a composite metric that combines two orthogonal criteria: information retention and throughput gains and finds that it allocates more resources to the most expressive layers compared to evolutionary search, specialized accelerators, or Shapley-value-based approaches that require expensive approximate inference.

A. Safronov · 1 citation · ⚡1
#small language model Preprint Aug 2026

HEPToolBench 1.2: Testing How Reliably Language Models Can Drive Particle Physics Software

HEPToolBench is introduced, a benchmark of 28 collider-simulation tasks scored by deterministic, task-specific scorers, plus a three-task structured-debugging extension, and moving syntax generation into deterministic software can substantially improve reliability for both small local and frontier models.

Unknown authors · 1 citation

Related blog posts

Microsoft Research Blog Aug 31, 2026

GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models

What if pathology foundation models could do more with less? GigaPath-Flash and GigaTIME-Flash cut computational demands while maintaining strong performance, opening the door to larger studies and broader exploration. The post GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models appeared first on Microsoft Research.